• 基于 hsv 的马赛克拼图效果 文章链接： 2018-5-30 更新内容 修复了生成失败中断的问题 不再读取数据库中的非图片文件 自定义最低相似标准 -far （默认1000） 修复命令行参数类型错误（竟然没人发现）。...
• The R1 company has recently bought a high rise building in the centre of Moscow for its main office. It’s time to decorate the new office, and the first thing to do is to write the company
Description  The R1 company has recently bought a high rise building in the centre of Moscow for its main office. It’s time to decorate the new office, and the first thing to do is to write the company’s slogan above the main entrance to the building.
The slogan of the company consists of n characters, so the decorators hung a large banner, n meters wide and 1 meter high, divided into n equal squares. The first character of the slogan must be in the first square (the leftmost) of the poster, the second character must be in the second square, and so on.
Of course, the R1 programmers want to write the slogan on the poster themselves. To do this, they have a large (and a very heavy) ladder which was put exactly opposite the k-th square of the poster. To draw the i-th character of the slogan on the poster, you need to climb the ladder, standing in front of the i-th square of the poster. This action (along with climbing up and down the ladder) takes one hour for a painter. The painter is not allowed to draw characters in the adjacent squares when the ladder is in front of the i-th square because the uncomfortable position of the ladder may make the characters untidy. Besides, the programmers can move the ladder. In one hour, they can move the ladder either a meter to the right or a meter to the left.
Drawing characters and moving the ladder is very tiring, so the programmers want to finish the job in as little time as possible. Develop for them an optimal poster painting plan!
Input  The first line contains two integers, n and k(1 ≤ k ≤ n ≤ 100) — the number of characters in the slogan and the initial position of the ladder, correspondingly. The next line contains the slogan as n characters written without spaces. Each character of the slogan is either a large English letter, or digit, or one of the characters: ‘.’, ‘!’, ‘,’, ‘?’.
Output  In t lines, print the actions the programmers need to make. In the i-th line print:
“LEFT” (without the quotes), if the i-th action was “move the ladder to the left”;  “RIGHT” (without the quotes), if the i-th action was “move the ladder to the right”;  “PRINTx” (without the quotes), if the i-th action was to “go up the ladder, paint character x, go down the ladder”.  The painting time (variable t) must be minimum possible. If there are multiple optimal painting plans, you can print any of them.
Sample Input  Input  2 2  R1  Output  PRINT 1  LEFT  PRINT R  Input  2 1  R1  Output  PRINT R  RIGHT  PRINT 1  Input  6 4  GO?GO!  Output  RIGHT  RIGHT  PRINT !  LEFT  PRINT O  LEFT  PRINT G  LEFT  PRINT ?  LEFT  PRINT O  LEFT  PRINT G  Hint  Note that the ladder cannot be shifted by less than one meter. The ladder can only stand in front of some square of the poster. For example, you cannot shift a ladder by half a meter and position it between two squares. Then go up and paint the first character and the second character.
题意：公司要为大标题上色，给出标题的长度，接着给出标题，移动梯子的代价是比较大的，所以要求找到最小的移动方案为标题的每个字上色。  思路：将梯子移动至较为靠近的一边，然后逐个输出即可。
#include <iostream>
#include <cstdio>
#include <cstring>
using namespace std;

int main()
{
int n,k;
cin>>n>>k;
char a[n+1];
for(int i=1;i<=n;i++)
cin>>a[i];
if(n>=2*k)
{
for(int i=k;i>1;i--)
{
cout<<"LEFT"<<endl;
}
for(int i=1;i<=n;i++)
{
if(i!=n)
{
printf("PRINT %c\nRIGHT\n",a[i]);
}
else
printf("PRINT %c\n",a[i]);
}
}
if(n<2*k)
{
for(int i=k;i<n;i++)
{
cout<<"RIGHT"<<endl;
}
for(int i=n;i>=1;i--)
{
if(i!=1)
{
printf("PRINT %c\nLEFT\n",a[i]);
}
else
printf("PRINT %c\n",a[i]);
}
}
return 0;
}
展开全文
• A developer tool for interacting with web services and other web resources that lets you make HTTP requests, set the entity body, and content type. This allows you to interact with web services and ...
• I have a background HTML5 video on my website :Your browser does not support the video element. Try this page in a modern browser!CSS :#video_background {position: absolute;bottom: 0px;right: 0px;min-...

I have a background HTML5 video on my website :
CSS :
#video_background {
position: absolute;
bottom: 0px;
right: 0px;
min-width: 100%;
min-height: 100%;
width: auto;
height: auto;
z-index: -1000;
overflow: hidden;
}
But the poster image doesn't work properly on mobile and tablet devices, it doesn't scale to fit the whole window resolution !!
My question is how I can make the poster image fit the window on any mobile / tablet device.

展开全文
• ICLR2019 poster papers 汇总

ICLR2019 Poster 文章汇总, 共478 papers

Convolutional Neural Networks on Non-uniform Geometrical Signals Using Euclidean Spectral Transformation Keywords:Non-uniform Fourier Transform, 3D Learning, CNN, surface reconstructionTL;DR:We use non-Euclidean Fourier Transformation of shapes defined by a simplicial complex for deep learning, achieving significantly better results than point-based sampling techiques used in current 3D learning literature. Augmented Cyclic Adversarial Learning for Low Resource Domain Adaptation Keywords:Domain adaptation, generative adversarial network, cyclic adversarial learning, speechTL;DR:A new cyclic adversarial learning augmented with auxiliary task model which improves domain adaptation performance in low resource supervised and unsupervised situations  Variance Networks: When Expectation Does Not Meet Your Expectations Keywords:deep learning, variational inference, variational dropoutTL;DR:It is possible to learn a zero-centered Gaussian distribution over the weights of a neural network by learning only variances, and it works surprisingly well. Initialized Equilibrium Propagation for Backprop-Free Training Keywords:credit assignment, energy-based models, biologically plausible learningTL;DR:We train a feedforward network without backprop by using an energy-based model to provide local targets Explaining Image Classifiers by Counterfactual Generation Keywords:Explainability, Interpretability, Generative Models, Saliency Map, Machine Learning, Deep LearningTL;DR:We compute saliency by using a strong generative model to efficiently marginalize over plausible alternative inputs, revealing concentrated pixel areas that preserve label information. SNIP: SINGLE-SHOT NETWORK PRUNING BASED ON CONNECTION SENSITIVITY Keywords:neural network pruning, connection sensitivityTL;DR:We present a new approach, SNIP, that is simple, versatile and interpretable; it prunes irrelevant connections for a given task at single-shot prior to training and is applicable to a variety of neural network models without modifications. Diagnosing and Enhancing VAE Models Keywords:variational autoencoder, generative modelsTL;DR:We closely analyze the VAE objective function and draw novel conclusions that lead to simple enhancements. Disjoint Mapping Network for Cross-modal Matching of Voices and Faces Keywords:NoneTL;DR:None Automatically Composing Representation Transformations as a Means for Generalization Keywords:compositionality, deep learning, metareasoningTL;DR:We explore the problem of compositional generalization and propose a means for endowing neural network architectures with the ability to compose themselves to solve these problems. Visual Reasoning by Progressive Module Networks Keywords:NoneTL;DR:None Bayesian Deep Convolutional Networks with Many Channels are Gaussian Processes Keywords:Deep Convolutional Neural Networks, Gaussian Processes, BayesianTL;DR:Finite-width SGD trained CNNs vs. infinitely wide fully Bayesian CNNs. Who wins? Learning to Learn without Forgetting by Maximizing Transfer and Minimizing Interference Keywords:NoneTL;DR:None Sparse Dictionary Learning by Dynamical Neural Networks Keywords:NoneTL;DR:None Eidetic 3D LSTM: A Model for Video Prediction and Beyond Keywords:NoneTL;DR:None ALISTA: Analytic Weights Are As Good As Learned Weights in LISTA Keywords:NoneTL;DR:None Three Mechanisms of Weight Decay Regularization Keywords:Generalization, Regularization, OptimizationTL;DR:We investigate weight decay regularization for different optimizers and identify three distinct mechanisms by which weight decay improves generalization. Learning Multimodal Graph-to-Graph Translation for Molecule Optimization Keywords:graph-to-graph translation, graph generation, molecular optimizationTL;DR:We introduce a graph-to-graph encoder-decoder framework for learning diverse graph translations. A Data-Driven and Distributed Approach to Sparse Signal Representation and Recovery Keywords:Sparsity, Compressive Sensing, Convolutional NetworkTL;DR:We use deep learning techniques to solve the sparse signal representation and recovery problem. On the Minimal Supervision for Training Any Binary Classifier from Only Unlabeled Data Keywords:learning from only unlabeled data, empirical risk minimization, unbiased risk estimatorTL;DR:Three class priors are all you need to train deep models from only U data, while any two should not be enough. Neural Logic Machines Keywords:Neural-Symbolic Computation, Rule Induction, First-Order LogicTL;DR:We propose the Neural Logic Machine (NLM), a neural-symbolic architecture for both inductive learning and logic reasoning. Neural Speed Reading with Structural-Jump-LSTM Keywords:natural language processing, speed reading, recurrent neural network, classificationTL;DR:We propose a new model for neural speed reading that utilizes the inherent punctuation structure of a text to define effective jumping and skipping behavior. Rigorous Agent Evaluation: An Adversarial Approach to Uncover Catastrophic Failures Keywords:agent evaluation, adversarial examples, robustness, safety, reinforcement learningTL;DR:We show that rare but catastrophic failures may be missed entirely by random testing, which poses issues for safe deployment. Our proposed approach for adversarial testing fixes this. Woulda, Coulda, Shoulda: Counterfactually-Guided Policy Search Keywords:NoneTL;DR:None signSGD via Zeroth-Order Oracle Keywords:nonconvex optimization, zeroth-order algorithm, black-box adversarial attackTL;DR:We design and analyze a new zeroth-order stochastic optimization algorithm, ZO-signSGD, and demonstrate its connection and application to black-box adversarial attacks in robust deep learning Preventing Posterior Collapse with delta-VAEs Keywords:Posterior Collapse, VAE, Autoregressive ModelsTL;DR: Avoid posterior collapse by lower bounding the rate. Algorithmic Framework for Model-based Deep Reinforcement Learning with Theoretical Guarantees Keywords:model-based reinforcement learning, sample efficiency, deep reinforcement learningTL;DR:We design model-based reinforcement learning algorithms with theoretical guarantees and achieve state-of-the-art results on Mujuco benchmark tasks when one million or fewer samples are permitted. Knowledge Flow: Improve Upon Your Teachers Keywords:Transfer Learning, Reinforcement LearningTL;DR:‘Knowledge Flow’ trains a deep net (student) by injecting information from multiple nets (teachers). The student is independent upon training and performs very well on learned tasks irrespective of the setting (reinforcement or supervised learning). Directed-Info GAIL: Learning Hierarchical Policies from Unsegmented Demonstrations using Directed Information Keywords:Imitation Learning, Reinforcement Learning, Deep LearningTL;DR:Learning Hierarchical Policies from Unsegmented Demonstrations using Directed Information A Max-Affine Spline Perspective of Recurrent Neural Networks Keywords:RNN, max-affine spline operatorsTL;DR:We provide new insights and interpretations of RNNs from a max-affine spline operators perspective. Learning to Navigate the Web Keywords:navigating web pages, reinforcement learning, q learning, curriculum learning, meta trainingTL;DR:We train reinforcement learning policies using reward augmentation, curriculum learning, and meta-learning to successfully navigate web pages. Training for Faster Adversarial Robustness Verification via Inducing ReLU Stability Keywords:verification, adversarial robustness, adversarial examples, stability, deep learning, regularizationTL;DR:We develop methods to train deep neural models that are both robust to adversarial perturbations and whose robustness is significantly easier to verify. Learning to Learn with Conditional Class Dependencies Keywords:meta-learning, learning to learn, few-shot learningTL;DR:CAML is an instance of MAML with conditional class dependencies. Hierarchical Visuomotor Control of Humanoids Keywords:hierarchical reinforcement learning, motor control, motion captureTL;DR:Solve tasks involving vision-guided humanoid locomotion, reusing locomotion behavior from motion capture data. Unsupervised Adversarial Image Reconstruction Keywords:NoneTL;DR:None Max-MIG: an Information Theoretic Approach for Joint Learning from Crowds Keywords:NoneTL;DR:None AutoLoss: Learning Discrete Schedule for Alternate Optimization Keywords:Meta Learning, AutoML, Optimization ScheduleTL;DR:We propose a unified formulation for iterative alternate optimization and develop AutoLoss, a framework to automatically learn and generate optimization schedules. Learning what and where to attend Keywords:Attention models, human feature importance, object recognition, cognitive scienceTL;DR:A large-scale dataset for training attention models for object recognition leads to more accurate, interpretable, and human-like object recognition. ROBUST ESTIMATION VIA GENERATIVE ADVERSARIAL NETWORKS Keywords:robust statistics, neural networks, minimax rate, data depth, contamination model, Tukey median, GANTL;DR:GANs are shown to provide us a new effective robust mean estimate against agnostic contaminations with both statistical optimality and practical tractability. INVASE: Instance-wise Variable Selection using Neural Networks Keywords:NoneTL;DR:None Meta-Learning with Latent Embedding Optimization Keywords:meta-learning, few-shot, miniImageNet, tieredImageNet, hypernetworks, generative, latent embedding, optimizationTL;DR:Latent Embedding Optimization (LEO) is a novel gradient-based meta-learner with state-of-the-art performance on the challenging 5-way 1-shot and 5-shot miniImageNet and tieredImageNet classification tasks. Non-vacuous Generalization Bounds at the ImageNet Scale: a PAC-Bayesian Compression Approach Keywords:generalization, deep-learning, pac-bayesTL;DR:We obtain non-vacuous generalization bounds on ImageNet-scale deep neural networks by combining an original PAC-Bayes bound and an off-the-shelf neural network compression method. Learning to Represent Edits Keywords:NoneTL;DR:None Neural Probabilistic Motor Primitives for Humanoid Control Keywords:Motor Primitives, Distillation, Reinforcement Learning, Continuous Control, Humanoid Control, Motion Capture, One-Shot ImitationTL;DR:Neural Probabilistic Motor Primitives compress motion capture tracking policies into one flexible model capable of one-shot imitation and reuse as a low-level controller. Differentiable Perturb-and-Parse: Semi-Supervised Parsing with a Structured Variational Autoencoder Keywords:differentiable dynamic programming, variational auto-encoder, dependency parsing, semi-supervised learningTL;DR:Differentiable dynamic programming over perturbed input weights with application to semi-supervised VAE Janossy Pooling: Learning Deep Permutation-Invariant Functions for Variable-Size Inputs Keywords:representation learning, permutation invariance, set functions, feature poolingTL;DR:We propose Janossy pooling, a method for learning deep permutation invariant functions designed to exploit relationships within the input sequence and tractable inference strategies such as a stochastic optimization procedure we call piSGD An Empirical Study of Example Forgetting during Deep Neural Network Learning Keywords:catastrophic forgetting, sample weighting, deep generalizationTL;DR:We show that catastrophic forgetting occurs within what is considered to be a single task and find that examples that are not prone to forgetting can be removed from the training set without loss of generalization. RNNs implicitly implement tensor-product representations Keywords:tensor-product representations, compositionality, neural network interpretability, recurrent neural networksTL;DR:RNNs implicitly implement tensor-product representations, a principled and interpretable method for representing symbolic structures in continuous space. Learning To Solve Circuit-SAT: An Unsupervised Differentiable Approach Keywords:Neuro-Symbolic Methods, Circuit Satisfiability, Neural SAT Solver, Graph Neural NetworksTL;DR:We propose a neural framework that can learn to solve the Circuit Satisfiability problem from (unlabeled) circuit instances. Dynamic Channel Pruning: Feature Boosting and Suppression Keywords:dynamic network, faster CNNs, channel pruningTL;DR:We make convolutional layers run faster by dynamically boosting and suppressing channels in feature computation. signSGD with Majority Vote is Communication Efficient and Fault Tolerant Keywords:large-scale learning, distributed systems, communication efficiency, convergence rate analysis, robust optimisationTL;DR:Workers send gradient signs to the server, and the update is decided by majority vote. We show that this algorithm is convergent, communication efficient and fault tolerant, both in theory and in practice. Bounce and Learn: Modeling Scene Dynamics with Real-World Bounces Keywords:NoneTL;DR:None K for the Price of 1: Parameter-efficient Multi-task and Transfer Learning Keywords:deep learning, mobile, transfer learning, multi-task learning, computer vision, small models, imagenet, inception, batch normalizationTL;DR:A novel and practically effective method to adapt pretrained neural networks to new tasks by retraining a minimal (e.g., less than 2%) number of parameters Towards Metamerism via Foveated Style Transfer Keywords:Metamerism, foveation, perception, style transfer, psychophysicsTL;DR:We introduce a novel feed-forward framework to generate visual metamers Post Selection Inference with Incomplete Maximum Mean Discrepancy Estimator Keywords:NoneTL;DR:None Emergent Coordination Through Competition Keywords:Multi-agent learning, Reinforcement LearningTL;DR:We introduce a new MuJoCo soccer environment for continuous multi-agent reinforcement learning research, and show that population-based training of independent reinforcement learners can learn cooperative behaviors Prior Convictions: Black-box Adversarial Attacks with Bandits and Priors Keywords:adversarial examples, gradient estimation, black-box attacks, model-based optimization, bandit optimizationTL;DR:We present a unifying view on black-box adversarial attacks as a gradient estimation problem, and then present a framework (based on bandits optimization) to integrate priors into gradient estimation, leading to significantly increased performance. Sample Efficient Imitation Learning for Continuous Control Keywords:Imitation Learning, Continuous Control, Reinforcement Learning, Inverse Reinforcement Learning, Conditional Generative Adversarial NetworkTL;DR:In this paper, we proposed a model-free, off-policy IL algorithm for continuous control. Experimental results showed that our algorithm achieves competitive results with GAIL while significantly reducing the environment interactions. Generative Code Modeling with Graphs Keywords:Generative Model, Source Code, Graph LearningTL;DR:Representing programs as graphs including semantics helps when generating programs Critical Learning Periods in Deep Networks Keywords:Critical Period, Deep Learning, Information Theory, Artificial Neuroscience, Information PlasticityTL;DR:Sensory deficits in early training phases can lead to irreversible performance loss in both artificial and neuronal networks, suggesting information phenomena as the common cause, and point to the importance of the initial transient and forgetting. CEM-RL: Combining evolutionary and gradient-based methods for policy search Keywords:evolution strategy, deep reinforcement learningTL;DR:We propose a new combination of evolution strategy and deep reinforcement learning which takes the best of both worlds LanczosNet: Multi-Scale Deep Graph Convolutional Networks Keywords:NoneTL;DR:None Excessive Invariance Causes Adversarial Vulnerability Keywords:Generalization, Adversarial Examples, Invariance, Information Theory, Invertible NetworksTL;DR:We show deep networks are not only too sensitive to task-irrelevant changes of their input, but also too invariant to a wide range of task-relevant changes, thus making vast regions in input space vulnerable to adversarial attacks. Hindsight policy gradients Keywords:reinforcement learning, policy gradients, multi-goal reinforcement learningTL;DR:We introduce the capacity to exploit information about the degree to which an arbitrary goal has been achieved while another goal was intended to policy gradient methods. Adaptive Gradient Methods with Dynamic Bound of Learning Rate Keywords:Optimization, SGD, Adam, GeneralizationTL;DR:Novel variants of optimization methods that combine the benefits of both adaptive and non-adaptive methods. Decoupled Weight Decay Regularization Keywords:NoneTL;DR:None Optimistic mirror descent in saddle-point problems: Going the extra (gradient) mile Keywords:Mirror descent, extra-gradient, generative adversarial networks, saddle-point problemsTL;DR:We show how the inclusion of an extra-gradient step in first-order GAN training methods can improve stability and lead to improved convergence results. DialogWAE: Multimodal Response Generation with Conditional Wasserstein Auto-Encoder Keywords:NoneTL;DR:None No Training Required: Exploring Random Encoders for Sentence Classification Keywords:NoneTL;DR:None Neural Graph Evolution: Automatic Robot Design Keywords:Reinforcement learning, graph neural networks, robotics, deep learning, transfer learningTL;DR:Automatic robotic design search with graph neural networks Function Space Particle Optimization for Bayesian Neural Networks Keywords:NoneTL;DR:None Structured Adversarial Attack: Towards General Implementation and Better Interpretability Keywords:NoneTL;DR:None Spherical CNNs on Unstructured Grids Keywords:Spherical CNN, unstructured grid, panoramic, semantic segmentation, parameter efficiencyTL;DR:We present a new CNN kernel for unstructured grids for spherical signals, and show significant accuracy and parameter efficiency gain on tasks such as 3D classfication and omnidirectional image segmentation. Optimal Transport Maps For Distribution Preserving Operations on Latent Spaces of Generative Models Keywords:generative models, optimal transport, distribution preserving operationsTL;DR:We propose a framework for modifying the latent space operations such that the distribution mismatch between the resulting outputs and the prior distribution the generative model was trained on is fully eliminated. Deep Lagrangian Networks: Using Physics as Model Prior for Deep Learning Keywords:Deep Model Learning, Robot ControlTL;DR:This paper introduces a physics prior for Deep Learning and applies the resulting network topology for model-based control. Accumulation Bit-Width Scaling For Ultra-Low Precision Training Of Deep Networks Keywords:reduced precision floating-point, partial sum accumulation bit-width, deep learning, trainingTL;DR:We present an analytical framework to determine accumulation bit-width requirements in all three deep learning training GEMMs and verify the validity and tightness of our method via benchmarking experiments. Deep Convolutional Networks as shallow Gaussian Processes Keywords:Gaussian process, CNN, ResNet, BayesianTL;DR:We show that CNNs and ResNets with appropriate priors on the parameters are Gaussian processes in the limit of infinitely many convolutional filters. Unsupervised Domain Adaptation for Distance Metric Learning Keywords:domain adaptation, distance metric learning, face recognitionTL;DR:A new theory of unsupervised domain adaptation for distance metric learning and its application to face recognition across diverse ethnicity variations. A comprehensive, application-oriented study of catastrophic forgetting in DNNs Keywords:incremental learning, deep neural networks, catatrophic forgetting, sequential learningTL;DR:We check DNN models for catastrophic forgetting using a new evaluation scheme that reflects typical application conditions, with surprising results. Posterior Attention Models for Sequence to Sequence Learning Keywords:posterior inference, attention, seq2seq learning, translationTL;DR:Computing attention based on posterior distribution leads to more meaningful attention and better performance Generative Question Answering: Learning to Answer the Whole Question Keywords:Question answering, question generation, reasoning, squad, clevrTL;DR:Question answering models that model the joint distribution of questions and answers can learn more than discriminative models Diversity and Depth in Per-Example Routing Models Keywords:conditional computation, routing models, depthTL;DR:Per-example routing models benefit from architectural diversity, but still struggle to scale to a large number of routing decisions. Selfless Sequential Learning Keywords:Lifelong learning, Continual Learning, Sequential learning, RegularizationTL;DR:A regularization strategy for improving the performance of sequential learning M^3RL: Mind-aware Multi-agent Management Reinforcement Learning Keywords:Multi-agent Reinforcement Learning, Deep Reinforcement LearningTL;DR:We propose Mind-aware Multi-agent Management Reinforcement Learning (M^3RL) for training a manager to motivate self-interested workers to achieve optimal collaboration by assigning suitable contracts to them. The Deep Weight Prior Keywords:deep learning, variational inference, prior distributionsTL;DR:The deep weight prior learns a generative model for kernels of convolutional neural networks, that acts as a prior distribution while training on new datasets. Efficient Multi-Objective Neural Architecture Search via Lamarckian Evolution Keywords:Neural Architecture Search, AutoML, AutoDL, Deep Learning, Evolutionary Algorithms, Multi-Objective OptimizationTL;DR:We propose a method for efficient Multi-Objective Neural Architecture Search based on Lamarckian inheritance and evolutionary algorithms. Quaternion Recurrent Neural Networks Keywords:NoneTL;DR:None Adversarial Audio Synthesis Keywords:audio, waveform, spectrogram, GAN, adversarial, WaveGAN, SpecGANTL;DR:Learning to synthesize raw waveform audio with GANs Preconditioner on Matrix Lie Group for SGD Keywords:preconditioner, stochastic gradient descent, Newton method, Fisher information, natural gradient, Lie groupTL;DR:We propose a new framework for preconditioner learning, derive new forms of preconditioners and learning methods, and reveal the relationship to methods like RMSProp, Adam, Adagrad, ESGD, KFAC, batch normalization, etc. Learning to Screen for Fast Softmax Inference on Large Vocabulary Neural Networks Keywords:NoneTL;DR:None Adaptive Posterior Learning: few-shot learning with a surprise-based memory module Keywords:metalearning, memory, few-shot, relational, self-attention, classification, sequential, reasoning, working memory, episodic memoryTL;DR:We introduce a model which generalizes quickly from few observations by storing surprising information and attending over the most relevant data at each time point. Probabilistic Planning with Sequential Monte Carlo methods Keywords:control as inference, probabilistic planning, sequential monte carlo, model based reinforcement learningTL;DR:Leveraging control as inference and Sequential Monte Carlo methods, we proposed a probabilistic planning algorithm. Plan Online, Learn Offline: Efficient Learning and Exploration via Model-Based Control Keywords:deep reinforcement learning, exploration, model-basedTL;DR:We propose a framework that incorporates planning for efficient exploration and learning in complex environments. DHER: Hindsight Experience Replay for Dynamic Goals Keywords:NoneTL;DR:None FlowQA: Grasping Flow in History for Conversational Machine Comprehension Keywords:Machine Comprehension, Conversational Agent, Natural Language Processing, Deep LearningTL;DR:We propose the Flow mechanism and an end-to-end architecture, FlowQA, that achieves SotA on two conversational QA datasets and a sequential instruction understanding task. Learning to Design RNA Keywords:matter engineering, bioinformatics, rna design, reinforcement learning, meta learning, neural architecture search, hyperparameter optimizationTL;DR:We learn to solve the RNA Design problem with reinforcement learning using meta learning and autoML approaches. Robust Conditional Generative Adversarial Networks Keywords:conditional GAN, unsupervised pathway, autoencoder, robustnessTL;DR:We introduce a new type of conditional GAN, which aims to leverage structure in the target space of the generator. We augment the generator with a new, unsupervised pathway to learn the target structure.  Top-Down Neural Model For Formulae Keywords:logic, formula, recursive neural networks, recurrent neural networksTL;DR:A top-down approach how to recursively represent propositional formulae by neural networks is presented. Cost-Sensitive Robustness against Adversarial Examples Keywords:Certified robustness, Adversarial examples, Cost-sensitive learningTL;DR:A general method for training certified cost-sensitive robust classifier against adversarial perturbations The role of over-parametrization in generalization of neural networks Keywords:Generalization, Over-Parametrization, Neural Networks, Deep LearningTL;DR:We suggest a generalization bound that could partly explain the improvement in generalization with over-parametrization. Diffusion Scattering Transforms on Graphs Keywords:graph neural networks, deep learning, stability, scattering transforms, convolutional neural networksTL;DR:Stability of scattering transform representations of graph data to deformations of the underlying graph support. Capsule Graph Neural Network Keywords:CapsNet, Graph embedding, GNNTL;DR:Inspired by CapsNet, we propose a novel architecture for graph embeddings on the basis of node features extracted from GNN. Energy-Constrained Compression for Deep Neural Networks via Weighted Sparse Projection and Layer Input Masking Keywords:NoneTL;DR:None Emerging Disentanglement in Auto-Encoder Based Unsupervised Image Content Transfer Keywords:Image-to-image Translation, Disentanglement, Autoencoders, FacesTL;DR:An image to image translation method which adds to one image the content of another thereby creating a new image. SGD Converges to Global Minimum in Deep Learning via Star-convex Path Keywords:NoneTL;DR:None Toward Understanding the Impact of Staleness in Distributed Machine Learning Keywords:NoneTL;DR:None Transfer Learning for Sequences via Learning to Collocate Keywords:transfer learning, recurrent neural network, attention, natural language processingTL;DR:Transfer learning for sequence via learning to align cell-level information across domains. Learning Procedural Abstractions and Evaluating Discrete Latent Temporal Structure Keywords:NoneTL;DR:None Unsupervised Speech Recognition via Segmental Empirical Output Distribution Matching Keywords:NoneTL;DR:None Adversarial Attacks on Graph Neural Networks via Meta Learning Keywords:graph mining, adversarial attacks, meta learning, graph neural networks, node classificationTL;DR:We use meta-gradients to attack the training procedure of deep neural networks for graphs. Maximal Divergence Sequential Autoencoder for Binary Software Vulnerability Detection Keywords:Vulnerabilities Detection, Sequential Auto-Encoder, Separable RepresentationTL;DR:We propose a novel method named Maximal Divergence Sequential Auto-Encoder that leverages Variational AutoEncoder representation for binary code vulnerability detection. Neural Program Repair by Jointly Learning to Localize and Repair Keywords:neural program repair, neural program embeddings, pointer networksTL;DR:Multi-headed Pointer Networks for jointly learning to localize and repair Variable Misuse bugs Information-Directed Exploration for Deep Reinforcement Learning Keywords:reinforcement learning, exploration, information directed samplingTL;DR:We develop a practical extension of Information-Directed Sampling for Reinforcement Learning, which accounts for parametric uncertainty and heteroscedasticity in the return distribution for exploration. Attention, Learn to Solve Routing Problems! Keywords:learning, routing problems, heuristics, attention, reinforce, travelling salesman problem, vehicle routing problem, orienteering problem, prize collecting travelling salesman problemTL;DR:Attention based model trained with REINFORCE with greedy rollout baseline to learn heuristics with competitive results on TSP and other routing problems L2-Nonexpansive Neural Networks Keywords:NoneTL;DR:None Improving Generalization and Stability of Generative Adversarial Networks Keywords:GAN, generalization, gradient penalty, zero centered, convergenceTL;DR:We propose a zero-centered gradient penalty for improving generalization and stability of GANs Adaptive Input Representations for Neural Language Modeling Keywords:Neural language modelingTL;DR:Variable capacity input word embeddings and SOTA on WikiText-103, Billion Word benchmarks. Neural Persistence: A Complexity Measure for Deep Neural Networks Using Algebraic Topology Keywords:Algebraic topology, persistent homology, network complexity, neural networkTL;DR:We develop a new topological complexity measure for deep neural networks and demonstrate that it captures their salient properties. Efficient Augmentation via Data Subsampling Keywords:data augmentation, invariance, subsampling, influenceTL;DR:Selectively augmenting difficult to classify points results in efficient training. Neural TTS Stylization with Adversarial and Collaborative Games Keywords:Text-To-Speech synthesis, GANsTL;DR:a generative adversarial network for style modeling in a text-to-speech system Optimal Control Via Neural Networks: A Convex Approach Keywords:NoneTL;DR:None CBOW Is Not All You Need: Combining CBOW with the Compositional Matrix Space Model Keywords:Text representation learning, Sentence embedding, Efficient training scheme, word2vecTL;DR:We present a novel training scheme for efficiently obtaining order-aware sentence representations. Stochastic Optimization of Sorting Networks via Continuous Relaxations Keywords:continuous relaxations, sorting, permutation, stochastic computation graphs, Plackett-LuceTL;DR:We provide a continuous relaxation to the sorting operator, enabling end-to-end, gradient-based stochastic optimization. Adaptivity of deep ReLU network for learning in Besov and mixed smooth Besov spaces: optimal rate and curse of dimensionality Keywords:NoneTL;DR:None Generating Multiple Objects at Spatially Distinct Locations Keywords:controllable image generation, text-to-image synthesis, generative model, generative adversarial network, ganTL;DR:Extend GAN architecture to obtain control over locations and identities of multiple objects within generated images. Near-Optimal Representation Learning for Hierarchical Reinforcement Learning Keywords:representation hierarchy reinforcement learningTL;DR:We translate a bound on sub-optimality of representations to a practical training objective in the context of hierarchical reinforcement learning. Understanding Composition of Word Embeddings via Tensor Decomposition Keywords:word embeddings, semantic composition, tensor decompositionTL;DR:We present a generative model for compositional word embeddings that captures syntactic relations, and provide empirical verification and evaluation. Structured Neural Summarization Keywords:Summarization, Graphs, Source CodeTL;DR:One simple trick to improve sequence models: Compose them with a graph model Graph Wavelet Neural Network Keywords:graph convolution, graph wavelet transform, graph Fourier transform, semi-supervised learningTL;DR:We present graph wavelet neural network (GWNN), a novel graph convolutional neural network (CNN), leveraging graph wavelet transform to address the shortcoming of previous spectral graph CNN methods that depend on graph Fourier transform. A rotation-equivariant convolutional neural network model of primary visual cortex Keywords:rotation equivariance, equivariance, primary visual cortex, V1, neuroscience, system identificationTL;DR:A rotation-equivariant CNN model of V1 that outperforms previous models and suggest functional groupings of V1 neurons. Supervised Community Detection with Line Graph Neural Networks Keywords:community detection, graph neural networks, belief propagation, energy landscape, non-backtracking matrixTL;DR:We propose a novel graph neural network architecture based on the non-backtracking matrix defined over the edge adjacencies and demonstrate its effectiveness in community detection tasks on graphs. Multiple-Attribute Text Rewriting Keywords:controllable text generation, generative models, conditional generative models, style transferTL;DR:A system for rewriting text conditioned on multiple controllable attributes Wasserstein Barycenter Model Ensembling Keywords:Wasserstein barycenter model ensemblingTL;DR:we propose to use Wasserstein barycenters for semantic model ensembling Policy Transfer with Strategy Optimization Keywords:transfer learning, reinforcement learning, modeling error, strategy optimizationTL;DR:We propose a policy transfer algorithm that can overcome large and challenging discrepancies in the system dynamics such as latency, actuator modeling error, etc. code2seq: Generating Sequences from Structured Representations of Code Keywords:source code, programs, code2seqTL;DR:We leverage the syntactic structure of source code to generate natural language sequences. Predict then Propagate: Graph Neural Networks meet Personalized PageRank Keywords:Graph, GCN, GNN, Neural network, Graph neural network, Message passing neural network, Semi-supervised classification, Semi-supervised learning, PageRank, Personalized PageRankTL;DR:Personalized propagation of neural predictions (PPNP) improves graph neural networks by separating them into prediction and propagation via personalized PageRank. Slimmable Neural Networks Keywords:Slimmable neural networks, mobile deep learning, accuracy-efficiency trade-offsTL;DR:We present a simple and general method to train a single neural network executable at different widths (number of channels in a layer), permitting instant and adaptive accuracy-efficiency trade-offs at runtime. Analysing Mathematical Reasoning Abilities of Neural Models Keywords:mathematics, dataset, algebraic, reasoningTL;DR:A dataset for testing mathematical reasoning (and algebraic generalization), and results on current sequence-to-sequence models. RotDCF: Decomposition of Convolutional Filters for Rotation-Equivariant Deep Networks Keywords:NoneTL;DR:None Execution-Guided Neural Program Synthesis Keywords:NoneTL;DR:None Dynamic Sparse Graph for Efficient Deep Learning Keywords:Sparsity, compression, training, accelerationTL;DR:We construct dynamic sparse graph via dimension-reduction search to reduce compute and memory cost in both DNN training and inference. Fixup Initialization: Residual Learning Without Normalization Keywords:deep learning, residual networks, initialization, batch normalization, layer normalizationTL;DR:All you need to train deep residual networks is a good initialization; normalization layers are not necessary. ProbGAN: Towards Probabilistic GAN with Theoretical Guarantees Keywords:Generative Adversarial Networks, Bayesian Deep Learning, Mode Collapse, Inception Score, Generator, Discriminator, CIFAR-10, STL-10, ImageNetTL;DR:A novel probabilistic treatment for GAN with theoretical guarantee. Exploration by random network distillation Keywords:reinforcement learning, exploration, curiosityTL;DR:A simple exploration bonus is introduced and achieves state of the art performance in 3 hard exploration Atari games. Unsupervised Learning of the Set of Local Maxima Keywords:NoneTL;DR:None On the Convergence of A Class of Adam-Type Algorithms for Non-Convex Optimization Keywords:nonconvex optimization, Adam, convergence analysisTL;DR:We analyze convergence of Adam-type algorithms and provide mild sufficient conditions to guarantee their convergence, we also show violating the conditions can makes an algorithm diverge. Minimum Divergence vs. Maximum Margin: an Empirical Comparison on Seq2Seq Models Keywords:NoneTL;DR:None GANSynth: Adversarial Neural Audio Synthesis Keywords:GAN, Audio, WaveNet, NSynth, MusicTL;DR:High-quality audio synthesis with GANs Sliced Wasserstein Auto-Encoders Keywords:optimal transport, Wasserstein distances, auto-encoders, unsupervised learningTL;DR:In this paper we use the sliced-Wasserstein distance to shape the latent distribution of an auto-encoder into any samplable prior distribution.  Learning Two-layer Neural Networks with Symmetric Inputs Keywords:Neural Network, Optimization, Symmetric Inputs, Moment-of-momentsTL;DR:We give an algorithm for learning a two-layer neural network with symmetric input distribution.  Learning to Understand Goal Specifications by Modelling Reward Keywords:instruction following, reward modelling, language understandingTL;DR:We propose AGILE, a framework for training agents to perform instructions from examples of respective goal-states. Do Deep Generative Models Know What They Don't Know? Keywords:NoneTL;DR:None Identifying and Controlling Important Neurons in Neural Machine Translation Keywords:neural machine translation, individual neurons, unsupervised, analysis, correlation, translation control, distributivity, localizationTL;DR:Unsupervised methods for finding, analyzing, and controlling important neurons in NMT Representing Formal Languages: A Comparison Between Finite Automata and Recurrent Neural Networks Keywords:Language recognition, Recurrent Neural Networks, Representation Learning, deterministic finite automaton, automatonTL;DR:Finite Automata Can be Linearly decoded from Language-Recognizing RNNs using low coarseness abstraction functions and high accuracy decoders.  Visual Explanation by Interpretation: Improving Visual Feedback Capabilities of Deep Neural Networks Keywords:model explanation, model interpretation, explainable ai, evaluationTL;DR:Interpretation by Identifying model-learned features that serve as indicators for the task of interest. Explain model decisions by highlighting the response of these features in test data. Evaluate explanations objectively with a controlled dataset. Don't let your Discriminator be fooled Keywords:GAN, generative models, computer visionTL;DR:A discriminator that is not easily fooled by adversarial example makes GAN training more robust and leads to a smoother objective. Latent Convolutional Models Keywords:latent models, convolutional networks, unsupervised learning, deep learning, modeling natural images, image restorationTL;DR:We present a new deep latent model of natural images that can be trained from unlabeled datasets and can be utilized to solve various image restoration tasks. A Universal Music Translation Network Keywords:NoneTL;DR:None How to train your MAML Keywords:meta-learning, deep-learning, few-shot learning, supervised learning, neural-networks, stochastic optimizationTL;DR:MAML is great, but it has many problems, we solve many of those problems and as a result we learn most hyper parameters end to end, speed-up training and inference and set a new SOTA in few-shot learning Learning a SAT Solver from Single-Bit Supervision Keywords:sat, search, graph neural network, theorem proving, proofTL;DR:We train a graph network to predict boolean satisfiability and show that it learns to search for solutions, and that the solutions it finds can be decoded from its activations. Learning Representations of Sets through Optimized Permutations Keywords:sets, representation learning, permutation invarianceTL;DR:Learn how to permute a set, then encode permuted set with RNN to obtain a set representation. Big-Little Net: An Efficient Multi-Scale Feature Representation for Visual and Speech Recognition Keywords:NoneTL;DR:None Unsupervised Hyper-alignment for Multilingual Word Embeddings Keywords:NoneTL;DR:None Visual Semantic Navigation using Scene Priors Keywords:NoneTL;DR:None NOODL: Provable Online Dictionary Learning and Sparse Coding Keywords:provable dictionary learning, sparse coding, support recovery, iterative hard thresholding, matrix factorization, neural architectures, noodlTL;DR:We present a provable algorithm for exactly recovering both factors of the dictionary learning model.  Stochastic Gradient/Mirror Descent: Minimax Optimality and Implicit Regularization Keywords:NoneTL;DR:None Active Learning with Partial Feedback Keywords:NoneTL;DR:None Gradient descent aligns the layers of deep linear networks Keywords:NoneTL;DR:None Data-Dependent Coresets for Compressing Neural Networks with Applications to Generalization Bounds Keywords:NoneTL;DR:None On the loss landscape of a class of deep neural networks with no bad local valleys Keywords:NoneTL;DR:None DOM-Q-NET: Grounded RL on Structured Language Keywords:Reinforcement Learning, Web Navigation, Graph Neural NetworksTL;DR:Graph-based Deep Q Network for Web Navigation  Boosting Robustness Certification of Neural Networks Keywords:Robustness certification, Adversarial Attacks, Abstract Interpretation, MILP Solvers, Verification of Neural NetworksTL;DR:We refine the over-approximation results from incomplete verifiers using MILP solvers to prove more robustness properties than state-of-the-art.  Learning To Simulate Keywords:Simulation in machine learning, reinforcement learning, policy gradients, image renderingTL;DR:We propose an algorithm that automatically adjusts parameters of a simulation engine to generate training data for a neural network such that validation accuracy is maximized. Towards Understanding Regularization in Batch Normalization Keywords:NoneTL;DR:None The Laplacian in RL: Learning Representations with Efficient Approximations Keywords:Laplacian, reinforcement learning, representationTL;DR:We propose a scalable method to approximate the eigenvectors of the Laplacian in the reinforcement learning context and we show that the learned representations can improve the performance of an RL agent. Predicting the Generalization Gap in Deep Networks with Margin Distributions Keywords:Deep learning, large margin, generalization bounds, generalization gap.TL;DR:We develop a new scheme to predict the generalization gap in deep networks with high accuracy. Adversarial Imitation via Variational Inverse Reinforcement Learning Keywords:Our method introduces the empowerment-regularized maximum-entropy inverse reinforcement learning to learn near-optimal rewards and policies from expert demonstrations.TL;DR:Inverse Reinforcement Learning, Imitation learning, Variational lnference, Learning from demonstrations Reasoning About Physical Interactions with Object-Oriented Prediction and Planning Keywords:structured scene representation, predictive models, intuitive physics, self-supervised learningTL;DR:We present a framework for learning object-centric representations suitable for planning in tasks that require an understanding of physics. LayoutGAN: Generating Graphic Layouts with Wireframe Discriminators Keywords:NoneTL;DR:None Learning Mixed-Curvature Representations in Product Spaces Keywords:embeddings, non-Euclidean geometry, manifolds, geometry of dataTL;DR:Product manifold embedding spaces with heterogenous curvature yield improved representations compared to traditional embedding spaces for a variety of structures. StrokeNet: A Neural Painting Environment Keywords:image generation, differentiable model, reinforcement learning, deep learning, model basedTL;DR:StrokeNet is a novel architecture where the agent is trained to draw by strokes on a differentiable simulation of the environment, which could effectively exploit the power of back-propagation. Harmonizing Maximum Likelihood with GANs for Multimodal Conditional Generation Keywords:conditional GANs, conditional image generation, multimodal generation, reconstruction loss, maximum likelihood estimation, moment matchingTL;DR:We prove that the mode collapse in conditional GANs is largely attributed to a mismatch between reconstruction loss and GAN loss and introduce a set of novel loss functions as alternatives for reconstruction loss. Measuring Compositionality in Representation Learning Keywords:compositionality, representation learning, evaluationTL;DR:This paper proposes a simple procedure for evaluating compositional structure in learned representations, and uses the procedure to explore the role of compositionality in four learning problems. Benchmarking Neural Network Robustness to Common Corruptions and Perturbations Keywords:robustness, benchmark, convnets, perturbationsTL;DR:We propose ImageNet-C to measure classifier corruption robustness and ImageNet-P to measure perturbation robustness ADef: an Iterative Algorithm to Construct Adversarial Deformations Keywords:Adversarial examples, deformations, deep neural networks, computer visionTL;DR:We propose a new, efficient algorithm to construct adversarial examples by means of deformations, rather than additive perturbations. Discriminator-Actor-Critic: Addressing Sample Inefficiency and Reward Bias in Adversarial Imitation Learning Keywords:deep learning, reinforcement learning, imitation learning, adversarial learningTL;DR:We address sample inefficiency and reward bias in adversarial imitation learning algorithms such as GAIL and AIRL. Doubly Reparameterized Gradient Estimators for Monte Carlo Objectives Keywords:variational autoencoder, reparameterization trick, IWAE, VAE, RWS, JVITL;DR:Doubly reparameterized gradient estimators provide unbiased variance reduction which leads to improved performance. Learning Recurrent Binary/Ternary Weights Keywords:Quantized Recurrent Neural Network, Hardware Implementation, Deep LearningTL;DR:We propose high-performance LSTMs with binary/ternary weights, that can greatly reduce implementation complexity Learning concise representations for regression by evolving networks of trees Keywords:regression, stochastic optimization, evolutionary compution, feature engineeringTL;DR:Representing the network architecture as a set of syntax trees and optimizing their structure leads to accurate and concise regression models.  Efficient Training on Very Large Corpora via Gramian Estimation Keywords:similarity learning, pairwise learning, matrix factorization, Gramian estimation, variance reduction, neural embedding models, recommender systemsTL;DR:We develop efficient methods to train neural embedding models with a dot-product structure, by reformulating the objective function in terms of generalized Gram matrices, and maintaining estimates of those matrices. MAE: Mutual Posterior-Divergence Regularization for Variational AutoEncoders Keywords:NoneTL;DR:None Residual Non-local Attention Networks for Image Restoration Keywords:Non-local network, attention network, image restoration, residual learningTL;DR:New state-of-the-art framework for image restoration Meta-Learning For Stochastic Gradient MCMC Keywords:Meta Learning, MCMCTL;DR:This paper proposes a method to automate the design of stochastic gradient MCMC proposal using meta learning approach.  Systematic Generalization: What Is Required and Can It Be Learned? Keywords:systematic generalization, language understanding, visual questions answering, neural module networksTL;DR:We show that modular structured models are the best in terms of systematic generalization and that their end-to-end versions don't generalize as well. Efficient Lifelong Learning with A-GEM Keywords:Lifelong Learning, Continual Learning, Catastrophic Forgetting, Few-shot TransferTL;DR:An efficient lifelong learning algorithm that provides a better trade-off between accuracy and time/ memory complexity compared to other algorithms.  Multi-step Retriever-Reader Interaction for Scalable Open-domain Question Answering Keywords:Open domain Question Answering, Reinforcement Learning, Query reformulationTL;DR:Paragraph retriever and machine reader interacts with each other via reinforcement learning to yield large improvements on open domain datasets Double Viterbi: Weight Encoding for High Compression Ratio and Fast On-Chip Reconstruction for Deep Neural Network Keywords:quantization, pruning, memory footprint, model compression, sparse matrixTL;DR:We present a new weight encoding scheme which enables high compression ratio and fast sparse-to-dense matrix conversion. Overcoming the Disentanglement vs Reconstruction Trade-off via Jacobian Supervision Keywords:disentangling, autoencoders, jacobian, face manipulationTL;DR:A method for learning image representations that are good for both disentangling factors of variation and obtaining faithful reconstructions. RotatE: Knowledge Graph Embedding by Relational Rotation in Complex Space Keywords:knowledge graph embedding, knowledge graph completion, adversarial samplingTL;DR:A new state-of-the-art approach for knowledge graph embedding. Guiding Policies with Language via Meta-Learning Keywords:meta-learning, language grounding, interactiveTL;DR:We propose a meta-learning method for interactively correcting policies with natural language. AdaShift: Decorrelation and Convergence of Adaptive Learning Rate Methods Keywords:optimizer, Adam, convergence, decorrelationTL;DR:We analysis and solve the non-convergence issue of Adam. AD-VAT: An Asymmetric Dueling mechanism for learning Visual Active Tracking Keywords:Active tracking, reinforcement learning, adversarial learning, multi agentTL;DR:We propose AD-VAT, where the tracker and the target object, viewed as two learnable agents, are opponents and can mutually enhance during training. Marginal Policy Gradients: A Unified Family of Estimators for Bounded Action Spaces with Applications Keywords:NoneTL;DR:None On Self Modulation for Generative Adversarial Networks Keywords:unsupervised learning, generative adversarial networks, deep generative modellingTL;DR:A simple GAN modification that improves performance across many losses, architectures, regularization schemes, and datasets.  Off-Policy Evaluation and Learning from Logged Bandit Feedback: Error Reduction via Surrogate Policy Keywords:NoneTL;DR:None Subgradient Descent Learns Orthogonal Dictionaries Keywords:Dictionary learning, Sparse coding, Non-convex optimization, TheoryTL;DR:Efficient dictionary learning by L1 minimization via a novel analysis of the non-convex non-smooth geometry. ClariNet: Parallel Wave Generation in End-to-End Text-to-Speech Keywords:NoneTL;DR:None MARGINALIZED AVERAGE ATTENTIONAL NETWORK FOR WEAKLY-SUPERVISED LEARNING Keywords:feature aggregation, weakly supervised learning, temporal action localizationTL;DR:A novel marginalized average attentional network for weakly-supervised temporal action localization  Towards GAN Benchmarks Which Require Generalization Keywords:evaluation, generative adversarial networks, adversarial divergencesTL;DR:We argue that GAN benchmarks must require a large sample from the model to penalize memorization and investigate whether neural network divergences have this property. A Closer Look at Few-shot Classification Keywords:few shot classification, meta-learningTL;DR: A detailed empirical study in few-shot classification that revealing challenges in standard evaluation setting and showing a new direction. Meta-Learning Probabilistic Inference for Prediction Keywords:probabilistic models, approximate inference, few-shot learning, meta-learningTL;DR:Novel framework for meta-learning that unifies and extends a broad class of existing few-shot learning methods. Achieves strong performance on few-shot learning benchmarks without requiring iterative test-time inference.  Deep reinforcement learning with relational inductive biases Keywords:relational reasoning, reinforcement learning, graph neural networks, starcraft, generalization, inductive biasTL;DR:Relational inductive biases improve out-of-distribution generalization capacities in model-free reinforcement learning agents Relaxed Quantization for Discretized Neural Networks Keywords:Quantization, Compression, Neural Networks, EfficiencyTL;DR:We introduce a technique that allows for gradient based training of quantized neural networks. Tree-Structured Recurrent Switching Linear Dynamical Systems for Multi-Scale Modeling Keywords:NoneTL;DR:None STCN: Stochastic Temporal Convolutional Networks Keywords:latent variables, variational inference, temporal convolutional networks, sequence modeling, auto-regressive modelingTL;DR:We combine the computational advantages of temporal convolutional architectures with the expressiveness of stochastic latent variables. Soft Q-Learning with Mutual-Information Regularization Keywords:NoneTL;DR:None On the Turing Completeness of Modern Neural Network Architectures Keywords:Transformer, NeuralGPU, Turing completenessTL;DR:We show that the Transformer architecture and the Neural GPU are Turing complete. Improving Differentiable Neural Computers Through Memory Masking, De-allocation, and Link Distribution Sharpness Control Keywords:NoneTL;DR:None Evaluating Robustness of Neural Networks with Mixed Integer Programming Keywords:verification, adversarial robustness, adversarial examples, deep learningTL;DR:We efficiently verify the robustness of deep neural models with over 100,000 ReLUs, certifying more samples than the state-of-the-art and finding more adversarial examples than a strong first-order attack. Random mesh projectors for inverse problems Keywords:imaging, inverse problems, subspace projections, random Delaunay triangulations, CNN, geophysics, regularizationTL;DR:We solve ill-posed inverse problems with scarce ground truth examples by estimating an ensemble of random projections of the model instead of the model itself. Multi-Agent Dual Learning Keywords:NoneTL;DR:None Complement Objective Training Keywords:optimization, entropy, image recognition, natural language understanding, adversarial attacks, deep learningTL;DR:We propose Complement Objective Training (COT), a new training paradigm that optimizes both the primary and complement objectives for effectively learning the parameters of neural networks. Mode Normalization Keywords:Deep Learning, Expert Models, Normalization, Computer VisionTL;DR:We present a novel normalization method for deep neural networks that is robust to multi-modalities in intermediate feature distributions. Detecting Egregious Responses in Neural Sequence-to-sequence Models Keywords:Deep Learning, Natural Language Processing, Adversarial Attacks, Dialogue Response GenerationTL;DR:This paper aims to provide an empirical answer to the question of whether well-trained dialogue response model can output malicious responses. Learning Actionable Representations with Goal Conditioned Policies Keywords:Representation Learning, Reinforcement LearningTL;DR:Learning state representations which capture factors necessary for control Verification of Non-Linear Specifications for Neural Networks Keywords:NoneTL;DR:None Generating Liquid Simulations with Deformation-aware Neural Networks Keywords:Learning weighting and deformations of space-time data sets for highly efficient approximations of liquid behavior.TL;DR:deformation learning, spatial transformer networks, fluid simulation DyRep: Learning Representations over Dynamic Graphs Keywords:Dynamic Graphs, Representation Learning, Dynamic Processes, Temporal Point Process, Attention, Latent RepresentationTL;DR:Models Representation Learning over dynamic graphs as latent hidden process bridging two observed processes of Topological Evolution of and Interactions on dynamic graphs. Trellis Networks for Sequence Modeling Keywords:sequence modeling, language modeling, recurrent networks, convolutional networks, trellis networksTL;DR:Trellis networks are a new sequence modeling architecture that bridges recurrent and convolutional models and sets a new state of the art on word- and character-level language modeling. Scalable Unbalanced Optimal Transport using Generative Adversarial Networks Keywords:unbalanced optimal transport, generative adversarial networks, population modelingTL;DR:We propose new methodology for unbalanced optimal transport using generative adversarial networks. Solving the Rubik's Cube with Approximate Policy Iteration Keywords:reinforcement learning, Rubik's Cube, approximate policy iteration, deep learning, deep reinforcement learningTL;DR:We solve the Rubik's Cube with pure reinforcement learning Variance Reduction for Reinforcement Learning in Input-Driven Environments Keywords:reinforcement learning, policy gradient, input-driven environments, variance reduction, baselineTL;DR:For environments dictated partially by external input processes, we derive an input-dependent baseline that provably reduces the variance for policy gradient methods and improves the policy performance in a wide range of RL tasks. Model-Predictive Policy Learning with Uncertainty Regularization for Driving in Dense Traffic Keywords:model-based reinforcement learning, stochastic video prediction, autonomous drivingTL;DR:A model-based RL approach which uses a differentiable uncertainty penalty to learn driving policies from purely observational data. GAN Dissection: Visualizing and Understanding Generative Adversarial Networks Keywords:GAN representations are examined in detail, and sets of representation units are found that control the generation of semantic concepts in the output.TL;DR:GANs, representation, interpretability, causality Improving MMD-GAN Training with Repulsive Loss Function Keywords:generative adversarial nets, loss function, maximum mean discrepancy, image generation, unsupervised learningTL;DR:Rearranging the terms in maximum mean discrepancy yields a much better loss function for the discriminator of generative adversarial nets Deterministic PAC-Bayesian generalization bounds for deep networks via generalizing noise-resilience Keywords:generalization, PAC-Bayes, SGD, learning theory, implicit regularizationTL;DR:We provide a PAC-Bayes based generalization guarantee for uncompressed, deterministic deep networks by generalizing noise-resilience of the network on the training data to the test data. Recall Traces: Backtracking Models for Efficient Reinforcement Learning Keywords:Model free RL, Variational InferenceTL;DR:A backward model of previous (state, action) given the next state, i.e. P(s_t, a_t | s_{t+1}), can be used to simulate additional trajectories terminating at states of interest! Improves RL learning efficiency. Stable Recurrent Models Keywords:stability, gradient descent, non-convex optimization, recurrent neural networksTL;DR:Stable recurrent models can be approximated by feed-forward networks and empirically perform as well as unstable models on benchmark tasks. The Limitations of Adversarial Training and the Blind-Spot Attack Keywords:Adversarial Examples, Adversarial Training, Blind-Spot AttackTL;DR:We show that even the strongest adversarial training methods cannot defend against adversarial examples crafted on slightly scaled and shifted test images. Efficiently testing local optimality and escaping saddles for ReLU networks Keywords:local optimality, second-order stationary point, escaping saddle points, nondifferentiability, ReLU, empirical riskTL;DR:A theoretical algorithm for testing local optimality and extracting descent directions at nondifferentiable points of empirical risks of one-hidden-layer ReLU networks. ProxylessNAS: Direct Neural Architecture Search on Target Task and Hardware Keywords:Neural Architecture Search, Efficient Neural NetworksTL;DR:Proxy-less neural architecture search for directly learning architectures on large-scale target task (ImageNet) while reducing the cost to the same level of normal training. Hierarchical Reinforcement Learning via Advantage-Weighted Information Maximization Keywords:Hierarchical reinforcement learning, Representation learning, Continuous controlTL;DR:This paper presents a hierarchical reinforcement learning framework based on deterministic option policies and mutual information maximization.  Generalizable Adversarial Training via Spectral Normalization Keywords:NoneTL;DR:None Adversarial Domain Adaptation for Stable Brain-Machine Interfaces Keywords:Brain-Machine Interfaces, Domain Adaptation, Adversarial NetworksTL;DR:We implement an adversarial domain adaptation network to stabilize a fixed Brain-Machine Interface against gradual changes in the recorded neural signals. Deep Online Learning Via Meta-Learning: Continual Adaptation for Model-Based RL Keywords:NoneTL;DR:None Deep Anomaly Detection with Outlier Exposure Keywords:confidence, uncertainty, anomaly, robustnessTL;DR:OE teaches anomaly detectors to learn heuristics for detecting unseen anomalies; experiments are in classification, density estimation, and calibration in NLP and vision settings; we do not tune on test distribution samples, unlike previous work Contingency-Aware Exploration in Reinforcement Learning Keywords:Reinforcement Learning, Exploration, Contingency-AwarenessTL;DR:We investigate contingency-awareness and controllable aspects in exploration and achieve state-of-the-art performance on Montezuma's Revenge without expert demonstrations. Context-adaptive Entropy Model for End-to-end Optimized Image Compression Keywords:image compression, deep learning, entropy modelTL;DR:Context-adaptive entropy model for use in end-to-end optimized image compression, which significantly improves compression performance Variational Discriminator Bottleneck: Improving Imitation Learning, Inverse RL, and GANs by Constraining Information Flow Keywords:reinforcement learning, generative adversarial networks, imitation learning, inverse reinforcement learning, information bottleneckTL;DR:Regularizing adversarial learning with an information bottleneck, applied to imitation learning, inverse reinforcement learning, and generative adversarial networks. Meta-learning with differentiable closed-form solvers Keywords:few-shot learning, one-shot learning, meta-learning, deep learning, ridge regression, classificationTL;DR:We propose a meta-learning approach for few-shot classification that achieves strong performance at high-speed by back-propagating through the solution of fast solvers, such as ridge regression or logistic regression. Learning Self-Imitating Diverse Policies Keywords:Reinforcement-learning, Imitation-learning, Ensemble-trainingTL;DR:Policy optimization by using past good rollouts from the agent; learning shaped rewards via divergence minimization; SVPG with JS-kernel for population-based exploration. ProxQuant: Quantized Neural Networks via Proximal Operators Keywords:Model quantization, Optimization, RegularizationTL;DR:A principled framework for model quantization using the proximal gradient method, with empirical evaluation and theoretical convergence analyses. Universal Transformers Keywords:sequence-to-sequence, rnn, transformer, machine translation, language understanding, learning to executeTL;DR:We introduce the Universal Transformer, a self-attentive parallel-in-time recurrent sequence model that outperforms Transformers and LSTMs on a wide range of sequence-to-sequence tasks, including machine translation. Learning to Adapt in Dynamic, Real-World Environments through Meta-Reinforcement Learning Keywords:meta-learning, reinforcement learning, meta reinforcement learning, online adaptationTL;DR:A model-based meta-RL algorithm that enables a real robot to adapt online in dynamic environments L-Shapley and C-Shapley: Efficient Model Interpretation for Structured Data Keywords:Model Interpretation, Feature SelectionTL;DR:We develop two linear-complexity algorithms for model-agnostic model interpretation based on the Shapley value, in the settings where the contribution of features to the target is well-approximated by a graph-structured factorization. Discovery of Natural Language Concepts in Individual Units of CNNs Keywords:interpretability of deep neural networks, natural language representationTL;DR:We show that individual units in CNN representations learned in NLP tasks are selectively responsive to natural language concepts. Towards the first adversarially robust neural network model on MNIST Keywords:NoneTL;DR:None Discriminator Rejection Sampling Keywords:GANs, rejection samplingTL;DR:We use a GAN discriminator to perform an approximate rejection sampling scheme on the output of the GAN generator. Harmonic Unpaired Image-to-image Translation Keywords:unpaired image-to-image translation, cyclegan, smoothness constraintTL;DR:Smooth regularization over sample graph for unpaired image-to-image translation results in significantly improved consistency Universal Successor Features Approximators Keywords:NoneTL;DR:None Gradient Descent Provably Optimizes Over-parameterized Neural Networks Keywords:theory, non-convex optimization, overparameterization, gradient descentTL;DR:We prove gradient descent achieves zero training loss with a linear rate on over-parameterized neural networks. Opportunistic Learning: Budgeted Cost-Sensitive Learning from Data Streams Keywords:Cost-Aware Learning, Feature Acquisition, Reinforcement Learning, Stream Learning, Deep Q-LearningTL;DR:An online algorithm for cost-aware feature acquisition and prediction DARTS: Differentiable Architecture Search Keywords:deep learning, autoML, neural architecture search, image classification, language modelingTL;DR:We propose a differentiable architecture search algorithm for both convolutional and recurrent networks, achieving competitive performance with the state of the art using orders of magnitude less computation resources. Feature-Wise Bias Amplification Keywords:NoneTL;DR:None The relativistic discriminator: a key element missing from standard GAN Keywords:Improving the quality and stability of GANs using a relativistic discriminator; IPM GANs (such as WGAN-GP) are a special case.TL;DR:AI, deep learning, generative models, GAN Understanding and Improving Interpolation in Autoencoders via an Adversarial Regularizer Keywords:We propose a regularizer that improves interpolation and autoencoders and show that it also improves the learned representation for downstream tasks.TL;DR:autoencoders, interpolation, unsupervised learning, representation learning, adversarial learning Quasi-hyperbolic momentum and Adam for deep learning Keywords:sgd, momentum, nesterov, adam, qhm, qhadam, optimizationTL;DR:Mix plain SGD and momentum (or do something similar with Adam) for great profit. Local SGD Converges Fast and Communicates Little Keywords:optimization, communication, theory, stochastic gradient descent, SGD, mini-batch, local SGD, parallel restart SGD, distributed trainingTL;DR:We prove that parallel local SGD achieves linear speedup with much lesser communication than parallel mini-batch SGD. Learning Finite State Representations of Recurrent Policy Networks Keywords:recurrent neural networks, finite state machine, quantization, interpretability, autoencoder, moore machine, reinforcement learning, imitation learning, representation, Atari, TomitaTL;DR:Extracting a finite state machine from a recurrent neural network via quantization for the purpose of interpretability with experiments on Atari. Multilingual Neural Machine Translation with Knowledge Distillation Keywords:NMT, Multilingual NMT, Knowledge DistillationTL;DR:We proposed a knowledge distillation based method to boost the accuracy of multilingual neural machine translation. MisGAN: Learning from Incomplete Data with Generative Adversarial Networks Keywords:generative models, missing dataTL;DR:This paper presents a GAN-based framework for learning the distribution from high-dimensional incomplete data. A Direct Approach to Robust Deep Learning Using Adversarial Networks Keywords:deep learning, adversarial learning, generative adversarial networksTL;DR:Jointly train an adversarial noise generating network with a classification network to provide better robustness to adversarial attacks. Combinatorial Attacks on Binarized Neural Networks Keywords:binarized neural networks, combinatorial optimization, integer programmingTL;DR:Gradient-based attacks on binarized neural networks are not effective due to the non-differentiability of such networks; Our IPROP algorithm solves this problem using integer optimization Exemplar Guided Unsupervised Image-to-Image Translation with Semantic Consistency Keywords:image-to-image translation, image generation, domain adaptationTL;DR:We propose the Exemplar Guided & Semantically Consistent Image-to-image Translation (EGSC-IT) network which conditions the translation process on an exemplar image in the target domain. ARM: Augment-REINFORCE-Merge Gradient for Stochastic Binary Networks Keywords:Antithetic sampling, variable augmentation, deep discrete latent variable models, variance reduction, variational auto-encoderTL;DR:An unbiased and low-variance gradient estimator for discrete latent variable models Building Dynamic Knowledge Graphs from Text using Machine Reading Comprehension Keywords:NoneTL;DR:None Information asymmetry in KL-regularized RL Keywords:Deep Reinforcement Learning, Continuous Control, RL as InferenceTL;DR:Limiting state information for the default policy can improvement performance, in a KL-regularized RL framework where both agent and default policy are optimized together TimbreTron: A WaveNet(CycleGAN(CQT(Audio))) Pipeline for Musical Timbre Transfer Keywords:Generative models, Timbre Transfer, Wavenet, CycleGANTL;DR:We present the TimbreTron, a pipeline for perfoming high-quality timbre transfer on musical waveforms using CQT-domain style transfer. Whitening and Coloring Batch Transform for GANs Keywords:NoneTL;DR:None Learnable Embedding Space for Efficient Neural Architecture Compression Keywords:Network Compression, Neural Architecture Search, Bayesian Optimization, Architecture EmbeddingTL;DR:We propose a method to incrementally learn an embedding space over the domain of network architectures, to enable the careful selection of architectures for evaluation during compressed architecture search. On the Sensitivity of Adversarial Robustness to Input Data Distributions Keywords:adversarial robustness, adversarial training, PGD training, adversarial perturbation, input data distributionTL;DR:Robustness performance of PGD trained models are sensitive to semantics-preserving transformation of image datasets, which implies the trickiness of evaluation of robust learning algorithms in practice. Minimal Images in Deep Neural Networks: Fragile Object Recognition in Natural Images Keywords:NoneTL;DR:None A Statistical Approach to Assessing Neural Network Robustness Keywords:neural network verification, multi-level splitting, formal verificationTL;DR:We introduce a statistical approach to assessing neural network robustness that provides an informative notion of how robust a network is, rather than just the conventional binary assertion of whether or not of property is violated. Improving Sequence-to-Sequence Learning via Optimal Transport Keywords:NoneTL;DR:None PATE-GAN: Generating Synthetic Data with Differential Privacy Guarantees Keywords:NoneTL;DR:None Integer Networks for Data Compression with Latent-Variable Models Keywords:data compression, variational models, network quantizationTL;DR:We train variational models with quantized networks for computational determinism. This enables using them for cross-platform data compression. Value Propagation Networks Keywords:Reinforcement Learning, Value Iteration, Navigation, Convolutional Neural Networks, Learning to planTL;DR:We present planners based on convnets that are sample-efficient and that generalize to larger instances of navigation and pathfinding problems. Bayesian Policy Optimization for Model Uncertainty Keywords:Bayes-Adaptive Markov Decision Process, Model Uncertainty, Bayes Policy OptimizationTL;DR:We formulate model uncertainty in Reinforcement Learning as a continuous Bayes-Adaptive Markov Decision Process and present a method for practical and scalable Bayesian policy optimization. Variational Bayesian Phylogenetic Inference Keywords:Bayesian phylogenetic inference, Variational inference, Subsplit Bayesian networksTL;DR:The first variational Bayes formulation of phylogenetic inference, a challenging inference problem over structures with intertwined discrete and continuous components LEARNING FACTORIZED REPRESENTATIONS FOR OPEN-SET DOMAIN ADAPTATION Keywords:NoneTL;DR:None On the Universal Approximability and Complexity Bounds of Quantized ReLU Neural Networks Keywords:Quantized Neural Networks, Universial Approximability, Complexity Bounds, Optimal Bit-widthTL;DR:This paper proves the universal approximability of quantized ReLU neural networks and puts forward the complexity bound given arbitrary error. Learning Localized Generative Models for 3D Point Clouds via Graph Convolution Keywords:A GAN using graph convolution operations with dynamically computed graphs from hidden featuresTL;DR:GAN, graph convolution, point clouds ACCELERATING NONCONVEX LEARNING VIA REPLICA EXCHANGE LANGEVIN DIFFUSION Keywords:NoneTL;DR:None Dynamically Unfolding Recurrent Restorer: A Moving Endpoint Control Method for Image Restoration Keywords:image restoration, differential equationTL;DR:We propose a novel method to handle image degradations of different levels by learning a diffusion terminal time. Our model can generalize to unseen degradation level and different noise statistic. Bias-Reduced Uncertainty Estimation for Deep Neural Classifiers Keywords:Uncertainty estimation, Deep learningTL;DR:We use snapshots from the training process to improve any uncertainty estimation method of a DNN classifier. CAMOU: Learning Physical Vehicle Camouflages to Adversarially Attack Detectors in the Wild Keywords:Adversarial Attack, Object Detection, Synthetic SimulationTL;DR:We propose a method to learn physical vehicle camouflage to adversarially attack object detectors in the wild. We find our camouflage effective and transferable. Learning Latent Superstructures in Variational Autoencoders for Deep Multidimensional Clustering Keywords:latent tree model, variational autoencoder, deep learning, latent variable model, bayesian network, structure learning, stepwise em, message passing, graphical model, multidimensional clustering, unsupervised learningTL;DR:We investigate a variant of variational autoencoders where there is a superstructure of discrete latent variables on top of the latent features. Learning Programmatically Structured Representations with Perceptor Gradients Keywords:NoneTL;DR:None Variational Autoencoders with Jointly Optimized Latent Dependency Structure Keywords:deep generative models, structure learningTL;DR:We propose a method for learning latent dependency structure in variational autoencoders. The Unusual Effectiveness of Averaging in GAN Training Keywords:NoneTL;DR:None Beyond Pixel Norm-Balls: Parametric Adversaries using an Analytically Differentiable Renderer Keywords:adversarial examples, norm-balls, differentiable rendererTL;DR:Enabled by a novel differentiable renderer, we propose a new metric that has real-world implications for evaluating adversarial machine learning algorithms, resolving the lack of realism of the existing metric based on pixel norms. Diversity is All You Need: Learning Skills without a Reward Function Keywords:reinforcement learning, unsupervised learning, skill discoveryTL;DR:We propose an algorithm for learning useful skills without a reward function, and show how these skills can be used to solve downstream tasks. Supervised Policy Update for Deep Reinforcement Learning Keywords:Deep Reinforcement LearningTL;DR:first posing and solving the sample efficiency optimization problem in the non-parameterized policy space, and then solving a supervised regression problem to find a parameterized policy that is near the optimal non-parameterized policy. Learning sparse relational transition models Keywords:Deictic reference, relational model, rule-based transition modelTL;DR:A new approach that learns a representation for describing transition models in complex uncertaindomains using relational rules.  Learning to Schedule Communication in Multi-agent Reinforcement Learning Keywords:NoneTL;DR:None Hierarchical RL Using an Ensemble of Proprioceptive Periodic Policies Keywords:NoneTL;DR:None Multi-class classification without multi-class labels Keywords:NoneTL;DR:None What do you learn from context? Probing for sentence structure in contextualized word representations Keywords:natural language processing, word embeddings, transfer learning, interpretabilityTL;DR:We probe for sentence structure in ELMo and related contextual embedding models. We find existing models efficiently encode syntax and show evidence of long-range dependencies, but only offer small improvements on semantic tasks. Spectral Inference Networks: Unifying Deep and Spectral Learning Keywords:spectral learning, unsupervised learning, manifold learning, dimensionality reductionTL;DR:We show how to learn spectral decompositions of linear operators with deep learning, and use it for unsupervised learning without a generative model. PeerNets: Exploiting Peer Wisdom Against Adversarial Attacks Keywords:NoneTL;DR:None Attentive Neural Processes Keywords:Neural Processes, Conditional Neural Processes, Stochastic Processes, Regression, AttentionTL;DR:A model for regression that learns conditional distributions of a stochastic process, by incorporating attention into Neural Processes. Representation Degeneration Problem in Training Natural Language Generation Models Keywords:NoneTL;DR:None Hierarchical interpretations for neural network predictions Keywords:interpretability, natural language processing, computer visionTL;DR:We introduce and validate hierarchical local interpretations, the first technique to automatically search for and display important interactions for individual predictions made by LSTMs and CNNs. Spreading vectors for similarity search Keywords:dimensionality reduction, similarity search, indexing, differential entropyTL;DR:We learn a neural network that uniformizes the input distribution, which leads to competitive indexing performance in high-dimensional space A Convergence Analysis of Gradient Descent for Deep Linear Neural Networks Keywords:Deep Learning, Learning Theory, Non-Convex OptimizationTL;DR:We analyze gradient descent for deep linear neural networks, providing a guarantee of convergence to global optimum at a linear rate. Feed-forward Propagation in Probabilistic Neural Networks with Categorical and Max Layers Keywords:probabilistic neural network, uncertainty, dropout, bayesian, softmax, argmax, logsumexpTL;DR:Approximating mean and variance of the NN output over noisy input / dropout / uncertain parameters. Analytic approximations for argmax, softmax and max layers. Measuring and regularizing networks in function space Keywords:function space, Hilbert space, empirical characterization, multitask learning, catastrophic forgetting, optimization, natural gradientTL;DR:We find movement in function space is not proportional to movement in parameter space during optimization. We propose a new natural-gradient style optimizer to address this. Fluctuation-dissipation relations for stochastic gradient descent Keywords:stochastic gradient descent, adaptive method, loss surface, HessianTL;DR:We prove fluctuation-dissipation relations for SGD, which can be used to (i) adaptively set learning rates and (ii) probe loss surfaces. Poincare Glove: Hyperbolic Word Embeddings Keywords:word embeddings, hyperbolic spaces, poincare ball, hypernymy, analogy, similarity, gaussian embeddingsTL;DR:We embed words in the hyperbolic space and make the connection with the Gaussian word embeddings. Episodic Curiosity through Reachability Keywords:deep learning, reinforcement learning, curiosity, exploration, episodic memoryTL;DR:We propose a novel model of curiosity based on episodic memory and the ideas of reachability which allows us to overcome the known "couch-potato" issues of prior work. Phase-Aware Speech Enhancement with Deep Complex U-Net Keywords:speech enhancement, deep learning, complex neural networks, phase estimationTL;DR:This paper proposes a novel complex masking method for speech enhancement along with a loss function for efficient phase estimation. Generative predecessor models for sample-efficient imitation learning Keywords:NoneTL;DR:None Adaptive Estimators Show Information Compression in Deep Neural Networks Keywords:deep neural networks, mutual information, information bottleneck, noise, L2 regularizationTL;DR:We developed robust mutual information estimates for DNNs and used them to observe compression in networks with non-saturating activation functions Multilingual Neural Machine Translation With Soft Decoupled Encoding Keywords:NoneTL;DR:None Approximating CNNs with Bag-of-local-Features models works surprisingly well on ImageNet Keywords:interpretability, representation learning, bag of features, deep learning, object recognitionTL;DR:Aggregating class evidence from many small image patches suffices to solve ImageNet, yields more interpretable models and can explain aspects of the decision-making of popular DNNs. Reward Constrained Policy Optimization Keywords:reinforcement learning, markov decision process, constrained markov decision process, deep learningTL;DR:For complex constraints in which it is not easy to estimate the gradient, we use the discounted penalty as a guiding signal. We prove that under certain assumptions it converges to a feasible solution. On the Relation Between the Sharpest Directions of DNN Loss and the SGD Step Length Keywords:optimization, generalization, theory of deep learning, SGD, hessianTL;DR:SGD is steered early on in training towards a region in which its step is too large compared to curvature, which impacts the rest of training.  Modeling the Long Term Future in Model-Based Reinforcement Learning Keywords:model-based reinforcement learning, variation inferenceTL;DR:incorporating, in the model, latent variables that encode future content improves the long-term prediction accuracy, which is critical for better planning in model-based RL. Understanding Straight-Through Estimator in Training Activation Quantized Neural Nets Keywords:straight-through estimator, quantized activation, binary neuronTL;DR:We make theoretical justification for the concept of straight-through estimator. DISTRIBUTIONAL CONCAVITY REGULARIZATION FOR GANS Keywords:NoneTL;DR:None LeMoNADe: Learned Motif and Neuronal Assembly Detection in calcium imaging videos Keywords:VAE, unsupervised learning, neuronal assemblies, calcium imaging analysisTL;DR:We present LeMoNADe, an end-to-end learned motif detection method directly operating on calcium imaging videos. Competitive experience replay Keywords:reinforcement learning, sparse reward, goal-based learningTL;DR:a novel method to learn with sparse reward using adversarial reward re-labeling Multi-Domain Adversarial Learning Keywords:multi-domain learning, domain adaptation, adversarial learning, H-divergence, deep representation learning, high-content microscopyTL;DR:Adversarial Domain adaptation and Multi-domain learning: a new loss to handle multi- and single-domain classes in the semi-supervised setting. ProMP: Proximal Meta-Policy Search Keywords:Meta-Reinforcement Learning, Meta-Learning, Reinforcement-LearningTL;DR:A novel and theoretically grounded meta-reinforcement learning algorithm Don't Settle for Average, Go for the Max: Fuzzy Sets and Max-Pooled Word Vectors Keywords:word vectors, sentence representations, distributed representations, fuzzy sets, bag-of-words, unsupervised learning, word vector compositionality, max-pooling, Jaccard indexTL;DR:Max-pooled word vectors with fuzzy Jaccard set similarity are an extremely competitive baseline for semantic similarity; we propose a simple dynamic variant that performs even better. Stable Opponent Shaping in Differentiable Games Keywords:multi-agent learning, multiple interacting losses, opponent shaping, exploitation, convergenceTL;DR:Opponent shaping is a powerful approach to multi-agent learning but can prevent convergence; our SOS algorithm fixes this with strong guarantees in all differentiable games. A Mean Field Theory of Batch Normalization Keywords:theory, batch normalization, mean field theory, trainabilityTL;DR:Batch normalization causes exploding gradients in vanilla feedforward networks. Learning Exploration Policies for Navigation Keywords:NoneTL;DR:None Distribution-Interpolation Trade off in Generative Models Keywords:generative models, latent distribution, Cauchy distribution, interpolationsTL;DR:We theoretically prove that linear interpolations are unsuitable for analysis of trained implicit generative models.  Learning to Describe Scenes with Programs Keywords:Structured scene representations, program synthesisTL;DR:We present scene programs, a structured scene representation that captures both low-level object appearance and high-level regularity in the scene. Visceral Machines: Risk-Aversion in Reinforcement Learning with Intrinsic Physiological Rewards Keywords:Reinforcement Learning, Simulation, Affective ComputingTL;DR:We present a novel approach to reinforcement learning that leverages a task-independent intrinsic reward function trained on peripheral pulse measurements that are correlated with human autonomic nervous system responses.  Deep Frank-Wolfe For Neural Network Optimization Keywords:optimization, conditional gradient, Frank-Wolfe, SVMTL;DR:We train neural networks by locally linearizing them and using a linear SVM solver (Frank-Wolfe) at each iteration. LEARNING TO PROPAGATE LABELS: TRANSDUCTIVE PROPAGATION NETWORK FOR FEW-SHOT LEARNING Keywords:few-shot learning, meta-learning, label propagation, manifold learningTL;DR:We propose a novel meta-learning framework for transductive inference that classifies the entire test set at once to alleviate the low-data problem. Improving the Generalization of Adversarial Training with Domain Adaptation Keywords:adversarial training, domain adaptation, adversarial example, deep learningTL;DR:We propose a novel adversarial training with domain adaptation method that significantly improves the generalization ability on adversarial examples from different attacks. Dimensionality Reduction for Representing the Knowledge of Probabilistic Models Keywords:metric learning, distance learning, dimensionality reduction, bound guaranteesTL;DR:dimensionality reduction for cases where examples can be represented as soft probability distributions Learning protein sequence embeddings using information from structure Keywords:sequence embedding, sequence alignment, RNN, LSTM, protein structure, amino acid sequence, contextual embeddings, transmembrane predictionTL;DR:We present a method for learning protein sequence embedding models using structural information in the form of global structural similarity between proteins and within protein residue-residue contacts. Variational Smoothing in Recurrent Neural Network Language Models Keywords:NoneTL;DR:None Biologically-Plausible Learning Algorithms Can Scale to Large Datasets Keywords:biologically plausible learning algorithm, ImageNet, sign-symmetry, feedback alignmentTL;DR:Biologically plausible learning algorithms, particularly sign-symmetry, work well on ImageNet Coarse-grain Fine-grain Coattention Network for Multi-evidence Question Answering Keywords:question answering, reading comprehension, nlp, natural language processing, attention, representation learningTL;DR:A new state-of-the-art model for multi-evidence question answering using coarse-grain fine-grain hierarchical attention. Learning a Meta-Solver for Syntax-Guided Program Synthesis Keywords:Syntax-guided Synthesis, Context Free Grammar, Logical Specification, Representation Learning, Meta Learning, Reinforcement LearningTL;DR:We propose a meta-learning framework that learns a transferable policy from only weak supervision to solve synthesis tasks with different logical specifications and grammars. Towards Robust, Locally Linear Deep Networks Keywords:robust derivatives, transparency, interpretabilityTL;DR:A scalable algorithm to establish robust derivatives of deep networks w.r.t. the inputs. How Important is a Neuron Keywords:NoneTL;DR:None Learning to Make Analogies by Contrasting Abstract Relational Structure Keywords:cognitive science, analogy, psychology, cognitive theory, cognition, abstraction, generalizationTL;DR:The most robust capacity for analogical reasoning is induced when networks learn analogies by contrasting abstract relational structures in their input domains. Learning what you can do before doing anything Keywords:unsupervised learning, vision, motion, action space, video prediction, variational modelsTL;DR:We learn a representation of an agent's action space from pure visual observations. We use a recurrent latent variable approach with a novel composability loss. Learning Grid Cells as Vector Representation of Self-Position Coupled with Matrix Representation of Self-Motion Keywords:NoneTL;DR:None Universal Stagewise Learning for Non-Convex Problems with Convergence on Averaged Solutions Keywords:NoneTL;DR:None Invariant and Equivariant Graph Networks Keywords:graph learning, equivariance, deep learningTL;DR:The paper provides a full characterization of permutation invariant and equivariant linear layers for graph data. Robustness May Be at Odds with Accuracy Keywords:adversarial examples, robust machine learning, robust optimization, deep feature representationsTL;DR:We show that adversarial robustness might come at the cost of standard classification performance, but also yields unexpected benefits. Feature Intertwiner for Object Detection Keywords:feature learning, computer vision, deep learningTL;DR:(Camera-ready version) A feature intertwiner module to leverage features from one accurate set to help the learning of another less reliable set. Adversarial Reprogramming of Neural Networks Keywords:Adversarial, Neural Networks, Machine Learning SecurityTL;DR:We introduce the first instance of adversarial attacks that reprogram the target model to perform a task chosen by the attacker---without the attacker needing to specify or compute the desired output for each test-time input. G-SGD: Optimizing ReLU Neural Networks in its Positively Scale-Invariant Space Keywords:NoneTL;DR:None From Hard to Soft: Understanding Deep Network Nonlinearities via Vector Quantization and Statistical Inference Keywords:Spline, Vector Quantization, Inference, Nonlinearities, Deep NetworkTL;DR:Reformulate deep networks nonlinearities from a vector quantization scope and bridge most known nonlinearities together. Aggregated Momentum: Stability Through Passive Damping Keywords:momentum, optimization, deep learning, neural networksTL;DR:We introduce a simple variant of momentum optimization which is able to outperform classical momentum, Nesterov, and Adam on deep learning tasks with minimal hyperparameter tuning. Variational Autoencoder with Arbitrary Conditioning Keywords:unsupervised learning, generative models, conditional variational autoencoder, variational autoencoder, missing features multiple imputation, inpaintingTL;DR:We propose an extension of conditional variational autoencoder that allows conditioning on an arbitrary subset of the features and sampling the remaining ones. Time-Agnostic Prediction: Predicting Predictable Video Frames Keywords:visual prediction, subgoal generation, bottleneck states, time-agnosticTL;DR:In visual prediction tasks, letting your predictive model choose which times to predict does two things: (i) improves prediction quality, and (ii) leads to semantically coherent "bottleneck state" predictions, which are useful for planning. A Closer Look at Deep Learning Heuristics: Learning rate restarts, Warmup and Distillation Keywords:deep learning heuristics, learning rate restarts, learning rate warmup, knowledge distillation, mode connectivity, SVCCATL;DR:We use empirical tools of mode connectivity and SVCCA to investigate neural network training heuristics of learning rate restarts, warmup and knowledge distillation. Self-Monitoring Navigation Agent via Auxiliary Progress Estimation Keywords:visual grounding, textual grounding, instruction-following, navigation agentTL;DR:We propose a self-monitoring agent for the Vision-and-Language Navigation task. Kernel Change-point Detection with Auxiliary Deep Generative Models Keywords:deep kernel learning, generative models, kernel two-sample test, time series change-point detectionTL;DR:In this paper, we propose KL-CPD, a novel kernel learning framework for time series CPD that optimizes a lower bound of test power via an auxiliary generative model as a surrogate to the abnormal distribution.  Unsupervised Learning via Meta-Learning Keywords:unsupervised learning, meta-learningTL;DR:An unsupervised learning method that uses meta-learning to enable efficient learning of downstream image classification tasks, outperforming state-of-the-art methods. Auxiliary Variational MCMC Keywords:NoneTL;DR:None Neural network gradient-based learning of black-box function interfaces Keywords:neural networks, black box functions, gradient descentTL;DR:Training DNNs to interface w\ black box functions w\o intermediate labels by using an estimator sub-network that can be replaced with the black box after training Self-Tuning Networks: Bilevel Optimization of Hyperparameters using Structured Best-Response Functions Keywords:hyperparameter optimization, game theory, optimizationTL;DR:We use a hypernetwork to predict optimal weights given hyperparameters, and jointly train everything together. Unsupervised Control Through Non-Parametric Discriminative Rewards Keywords:deep reinforcement learning, goals, UVFA, mutual informationTL;DR:Unsupervised reinforcement learning method for learning a policy to robustly achieve perceptually specified goals. Interpolation-Prediction Networks for Irregularly Sampled Time Series Keywords:irregular sampling, multivariate time series, supervised learning, interpolation, missing dataTL;DR:This paper presents a new deep learning architecture for addressing the problem of supervised learning with sparse and irregularly sampled multivariate time series. Riemannian Adaptive Optimization Methods Keywords:Riemannian optimization, adaptive, hyperbolic, curvature, manifold, adam, amsgrad, adagrad, rsgd, convergenceTL;DR:Adapting Adam, Amsgrad, Adagrad to Riemannian manifolds.  Minimal Random Code Learning: Getting Bits Back from Compressed Model Parameters Keywords:compression, neural networks, bits-back argument, Bayesian, Shannon, information theoryTL;DR:This paper proposes an effective method to compress neural networks based on recent results in information theory. Characterizing Audio Adversarial Examples Using Temporal Dependency Keywords:audio adversarial example, mitigation, detection, machine learningTL;DR:Adversarial audio discrimination using temporal dependency Equi-normalization of Neural Networks Keywords:convolutional neural networks, Normalization, Sinkhorn, RegularizationTL;DR:Fast iterative algorithm to balance the energy of a network while staying in the same functional equivalence class Generalized Tensor Models for Recurrent Neural Networks Keywords:expressive power, recurrent neural networks, Tensor-Train decompositionTL;DR:Analysis of expressivity and generality of recurrent neural networks with ReLu nonlinearities using Tensor-Train decomposition. Wizard of Wikipedia: Knowledge-Powered Conversational Agents Keywords:dialogue, knowledge, language, conversationTL;DR:We build knowledgeable conversational agents by conditioning on Wikipedia + a new supervised task. Are adversarial examples inevitable? Keywords:adversarial examples, neural networks, securityTL;DR:This paper identifies classes of problems for which adversarial examples are inescapable, and derives fundamental bounds on the susceptibility of any classifier to adversarial examples.  A Variational Inequality Perspective on Generative Adversarial Networks Keywords:optimization, variational inequality, games, saddle point, extrapolation, averaging, extragradient, generative modeling, generative adversarial networkTL;DR:We cast GANs in the variational inequality framework and import techniques from this literature to optimize GANs better; we give algorithmic extensions and empirically test their performance for training GANs. Learning-Based Frequency Estimation Algorithms Keywords:streaming algorithms, heavy-hitters, Count-Min, Count-SketchTL;DR:Data stream algorithms can be improved using deep learning, while retaining performance guarantees. From Language to Goals: Inverse Reinforcement Learning for Vision-Based Instruction Following Keywords:inverse reinforcement learning, language grounding, instruction following, language-based learningTL;DR:We ground language commands in a high-dimensional visual environment by learning language-conditioned rewards using inverse reinforcement learning. Backpropamine: training self-modifying neural networks with differentiable neuromodulated plasticity Keywords:meta-learning, reinforcement learning, plasticity, neuromodulation, Hebbian learning, recurrent neural networksTL;DR:Neural networks can be trained to modify their own connectivity, improving their online learning performance on challenging tasks. Recurrent Experience Replay in Distributed Reinforcement Learning Keywords:RNN, LSTM, experience replay, distributed training, reinforcement learningTL;DR:Investigation on combining recurrent neural networks and experience replay leading to state-of-the-art agent on both Atari-57 and DMLab-30 using single set of hyper-parameters. A Generative Model For Electron Paths Keywords:Molecules, Reaction Prediction, Graph Neural Networks, Deep Generative ModelsTL;DR:A generative model for reaction prediction that learns the mechanistic electron steps of a reaction directly from raw reaction data. Modeling Uncertainty with Hedged Instance Embeddings Keywords:uncertainty, instance embedding, metric learning, probabilistic embeddingTL;DR:The paper proposes using probability distributions instead of points for instance embeddings tasks such as recognition and verification. Beyond Greedy Ranking: Slate Optimization via List-CVAE Keywords:CVAE, VAE, recommendation system, slate optimization, whole page optimizationTL;DR:We used a CVAE type model structure to learn to directly generate slates/whole pages for recommendation systems. Stochastic Prediction of Multi-Agent Interactions from Partial Observations Keywords:Dynamics modeling, partial observations, multi-agent interactions, predictive modelsTL;DR:We present a method which learns to integrate temporal information and ambiguous visual information in the context of interacting agents. GamePad: A Learning Environment for Theorem Proving Keywords:Theorem proving, ITP, systems, neural embeddingsTL;DR:We introduce a system called GamePad to explore the application of machine learning methods to theorem proving in the Coq proof assistant. GLUE: A Multi-Task Benchmark and Analysis Platform for Natural Language Understanding Keywords:natural language understanding, multi-task learning, evaluationTL;DR:We present a multi-task benchmark and analysis platform for evaluating generalization in natural language understanding systems. On Computation and Generalization of Generative Adversarial Networks under Spectrum Control Keywords:NoneTL;DR:None Large-Scale Study of Curiosity-Driven Learning Keywords:exploration, curiosity, intrinsic reward, no extrinsic reward, unsupervised, no-reward, skillsTL;DR:An agent trained only with curiosity, and no extrinsic reward, does surprisingly well on 54 popular environments, including the suite of Atari games, Mario etc. Unsupervised Discovery of Parts, Structure, and Dynamics Keywords:Self-Supervised Learning, Visual Prediction, Hierarchical ModelsTL;DR:Learning object parts, hierarchical structure, and dynamics by watching how they move Music Transformer: Generating Music with Long-Term Structure Keywords:music generationTL;DR:We show the first successful use of Transformer in generating music that exhibits long-term structure.  BabyAI: A Platform to Study the Sample Efficiency of Grounded Language Learning Keywords:language, learning, efficiency, imitation learning, reinforcement learningTL;DR:We present the BabyAI platform for studying data efficiency of language learning with a human in the loop Analyzing Inverse Problems with Invertible Neural Networks Keywords:Inverse problems, Neural Networks, Uncertainty, Invertible Neural NetworksTL;DR:To analyze inverse problems with Invertible Neural Networks RelGAN: Relational Generative Adversarial Networks for Text Generation Keywords:NoneTL;DR:None The Singular Values of Convolutional Layers Keywords:singular values, operator norm, convolutional layers, regularizationTL;DR:We characterize the singular values of the linear transformation associated with a standard 2D multi-channel convolutional layer, enabling their efficient computation.  An Empirical study of Binary Neural Networks' Optimisation Keywords:NoneTL;DR:None Approximability of Discriminators Implies Diversity in GANs Keywords:Theory, Generative adversarial networks, Mode collapse, GeneralizationTL;DR:GANs can in principle learn distributions sample-efficiently, if the discriminator class is compact and has strong distinguishing power against the particular generator class. Learning Embeddings into Entropic Wasserstein Spaces Keywords:Embedding, Wasserstein, Sinkhorn, Optimal TransportTL;DR:We show that Wasserstein spaces are good targets for embedding data with complex semantic structure. DeepOBS: A Deep Learning Optimizer Benchmark Suite Keywords:deep learning, optimizationTL;DR:We provide a software package that drastically simplifies, automates, and improves the evaluation of deep learning optimizers. InfoBot: Transfer and Exploration via the Information Bottleneck Keywords:Information bottleneck, policy transfer, policy generalization, explorationTL;DR:Training agents with goal-policy information bottlenecks promotes transfer and yields a powerful exploration bonus The Comparative Power of ReLU Networks and Polynomial Kernels in the Presence of Sparse Latent Structure Keywords:theory, representational power, universal approximators, polynomial kernels, latent sparsity, beyond worst case, separation resultTL;DR:Beyond-worst-case analysis of the representational power of ReLU nets & polynomial kernels -- in particular in the presence of sparse latent structure. Learning Implicitly Recurrent CNNs Through Parameter Sharing Keywords:deep learning, architecture search, computer visionTL;DR:We propose a method that enables CNN folding to create recurrent connections Learning Particle Dynamics for Manipulating Rigid Bodies, Deformable Objects, and Fluids Keywords:Dynamics modeling, Control, Particle-Based RepresentationTL;DR:Learning particle dynamics with dynamic interaction graphs for simulating and control rigid bodies, deformable objects, and fluids.  Regularized Learning for Domain Adaptation under Label Shifts Keywords:Deep Learning, Domain Adaptation, Label Shift, Importance Weights, GeneralizationTL;DR:A practical and provably guaranteed approach for training efficiently classifiers in the presence of label shifts between Source and Target data sets Von Mises-Fisher Loss for Training Sequence to Sequence Models with Continuous Outputs Keywords:Language Generation, Regression, Word Embeddings, Machine TranslationTL;DR:Language generation using seq2seq models which produce word embeddings instead of a softmax based distribution over the vocabulary at each step enabling much faster training while maintaining generation quality Relational Forward Models for Multi-Agent Learning Keywords:multi-agent reinforcement learning, relational reasoning, forward modelsTL;DR:Relational Forward Models for multi-agent learning make accurate predictions of agents' future behavior, they produce intepretable representations and can be used inside agents. Imposing Category Trees Onto Word-Embeddings Using A Geometric Construction Keywords:category tree, word-embeddings, geometryTL;DR:we show a geometric method to perfectly encode categroy tree information into pre-trained word-embeddings. Two-Timescale Networks for Nonlinear Value Function Approximation Keywords:Reinforcement learning, policy evaluation, nonlinear function approximationTL;DR:We propose an architecture for learning value functions which allows the use of any linear policy evaluation algorithm in tandem with nonlinear feature learning. Diversity-Sensitive Conditional Generative Adversarial Networks Keywords:Conditional Generative Adversarial Network, mode-collapse, multi-modal generation, image-to-image translation, image in-painting, video predictionTL;DR:We propose a simple and general approach that avoids a mode collapse problem in various conditional GANs. Query-Efficient Hard-label Black-box Attack: An Optimization-based Approach Keywords:NoneTL;DR:None Rethinking the Value of Network Pruning Keywords:In structured network pruning, fine-tuning a pruned model only gives comparable performance with training it from scratch.TL;DR:network pruning, network compression, architecture search, train from scratch Hyperbolic Attention Networks Keywords:Hyperbolic Geometry, Attention Methods, Reasoning on Graphs, Relation Learning, Scale Free Graphs, Transformers, Power LawTL;DR:We propose to incorporate inductive biases and operations coming from hyperbolic geometry to improve the attention mechanism of the neural networks. Learning from Positive and Unlabeled Data with a Selection Bias Keywords:NoneTL;DR:None Adv-BNN: Improved Adversarial Defense through Robust Bayesian Neural Network Keywords:NoneTL;DR:None Optimal Completion Distillation for Sequence Learning Keywords:Sequence Learning, Edit Distance, Speech Recognition, Deep Reinforcement LearningTL;DR:Optimal Completion Distillation (OCD) is a training procedure for optimizing sequence to sequence models based on edit distance which achieves state-of-the-art on end-to-end Speech Recognition tasks. Caveats for information bottleneck in deterministic scenarios Keywords:Information bottleneck behaves in surprising ways whenever the output is a deterministic function of the input.TL;DR:information bottleneck, supervised learning, deep learning, information theory Deep Learning 3D Shapes Using Alt-az Anisotropic 2-Sphere Convolution Keywords:Spherical Convolution, Geometric deep learning, 3D shape analysisTL;DR:A method for applying deep learning to 3D surfaces using their spherical descriptors and alt-az anisotropic convolution on 2-sphere. Small nonlinearities in activation functions create bad local minima in neural networks Keywords:spurious local minima, loss surface, optimization landscape, neural networkTL;DR:We constructively prove that even the slightest nonlinear activation functions introduce spurious local minima, for general datasets and activation functions. Information Theoretic lower bounds on negative log likelihood Keywords:latent variable modeling, rate-distortion theory, log likelihood boundsTL;DR:Use rate-distortion theory to bound how much a latent variable model can be improved Preferences Implicit in the State of the World Keywords:Preference learning, Inverse reinforcement learning, Inverse optimal stochastic control, Maximum entropy reinforcement learning, Apprenticeship learningTL;DR:When a robot is deployed in an environment that humans have been acting in, the state of the environment is already optimized for what humans want, and we can use this to infer human preferences. A Kernel Random Matrix-Based Approach for Sparse PCA Keywords:NoneTL;DR:None Bayesian Prediction of Future Street Scenes using Synthetic Likelihoods Keywords:bayesian inference, segmentation, anticipation, multi-modalityTL;DR:Dropout based Bayesian inference is extended to deal with multi-modality and is evaluated on scene anticipation tasks. There Are Many Consistent Explanations of Unlabeled Data: Why You Should Average Keywords:semi-supervised learning, computer vision, classification, consistency regularization, flatness, weight averaging, stochastic weight averagingTL;DR:Consistency-based models for semi-supervised learning do not converge to a single point but continue to explore a diverse set of plausible solutions on the perimeter of a flat region. Weight averaging helps improve generalization performance. Large-Scale Answerer in Questioner's Mind for Visual Dialog Question Generation Keywords:NoneTL;DR:None Graph HyperNetworks for Neural Architecture Search Keywords:NoneTL;DR:None DELTA: DEEP LEARNING TRANSFER USING FEATURE MAP WITH ATTENTION FOR CONVOLUTIONAL NETWORKS Keywords:transfer learning, deep learning, regularization, attention, cnnTL;DR:improving deep transfer learning with regularization using attention based feature maps textTOvec: DEEP CONTEXTUALIZED NEURAL AUTOREGRESSIVE TOPIC MODELS OF LANGUAGE WITH DISTRIBUTED COMPOSITIONAL PRIOR Keywords:neural topic model, natural language processing, text representation, language modeling, information retrieval, deep learningTL;DR:Unified neural model of topic and language modeling to introduce language structure in topic models for contextualized topic vectors  Amortized Bayesian Meta-Learning Keywords:variational inference, meta-learning, few-shot learning, uncertainty quantificationTL;DR:We propose a meta-learning method which efficiently amortizes hierarchical variational inference across training episodes. Probabilistic Recursive Reasoning for Multi-Agent Reinforcement Learning Keywords:Multi-agent Reinforcement Learning, Recursive ReasoningTL;DR:We proposed a novel probabilisitic recursive reasoning (PR2) framework for multi-agent deep reinforcement learning tasks. Learning Neural PDE Solvers with Convergence Guarantees Keywords:Partial differential equation, deep learningTL;DR:We learn a fast neural solver for PDEs that has convergence guarantees. A new dog learns old tricks: RL finds classic optimization algorithms Keywords:reinforcement learning, algorithms, adwords, knapsack, secretaryTL;DR:By combining ideas from traditional algorithms design and reinforcement learning, we introduce a novel framework for learning algorithms that solve online combinatorial optimization problems. Deep Graph Infomax Keywords:Unsupervised Learning, Graph Neural Networks, Graph Convolutions, Mutual Information, Infomax, Deep LearningTL;DR:A new method for unsupervised representation learning on graphs, relying on maximizing mutual information between local and global representations in a graph. State-of-the-art results, competitive with supervised learning. Theoretical Analysis of Auto Rate-Tuning by Batch Normalization Keywords:batch normalization, scale invariance, learning rate, stationary pointTL;DR:We give a theoretical analysis of the ability of batch normalization to automatically tune learning rates, in the context of finding stationary points for a deep learning objective. Per-Tensor Fixed-Point Quantization of the Back-Propagation Algorithm Keywords:deep learning, reduced precision, fixed-point, quantization, back-propagation algorithmTL;DR:We analyze and determine the precision requirements for training neural networks when all tensors, including back-propagated signals and weight accumulators, are quantized to fixed-point format. FUNCTIONAL VARIATIONAL BAYESIAN NEURAL NETWORKS Keywords:functional variational inference, Bayesian neural networks, stochastic processesTL;DR:We perform functional variational inference on the stochastic processes defined by Bayesian neural networks. NADPEx: An on-policy temporally consistent exploration method for deep reinforcement learning Keywords:NoneTL;DR:None SPIGAN: Privileged Adversarial Learning from Simulation Keywords:domain adaptation, GAN, semantic segmentation, simulation, privileged informationTL;DR:An unsupervised sim-to-real domain adaptation method for semantic segmentation using privileged information from a simulator with GAN-based image translation. Generating Multi-Agent Trajectories using Programmatic Weak Supervision Keywords:deep learning, generative models, imitation learning, hierarchical methods, data programming, weak supervision, spatiotemporalTL;DR:We blend deep generative models with programmatic weak supervision to generate coordinated multi-agent trajectories of significantly higher quality than previous baselines. Label super-resolution networks Keywords:weakly supervised segmentation, land cover mapping, medical imagingTL;DR:Super-resolving coarse labels into pixel-level labels, applied to aerial imagery and medical scans. ANYTIME MINIBATCH: EXPLOITING STRAGGLERS IN ONLINE DISTRIBUTED OPTIMIZATION Keywords:distributed optimization, gradient descent, minibatch, stragglersTL;DR:Accelerate distributed optimization by exploiting stragglers. Sample Efficient Adaptive Text-to-Speech Keywords:few shot, meta learning, text to speech, wavenetTL;DR:Sample efficient algorithms to adapt a text-to-speech model to a new voice style with the state-of-the-art performance. Practical lossless compression with latent variables using bits back coding Keywords:compression, variational auto-encoders, deep latent gaussian models, lossless compression, latent variables, approximate inference, variational inferenceTL;DR:We do lossless compression of large image datasets using a VAE, beat existing compression algorithms. Kernel RNN Learning (KeRNL) Keywords:RNNs, Biologically plausible learning rules, Algorithm, Neural Networks, Supervised LearningTL;DR:A biologically plausible learning rule for training recurrent neural networks Deep, Skinny Neural Networks are not Universal Approximators Keywords:This paper proves that skinny neural networks cannot approximate certain functions, no matter how deep they are.TL;DR:neural network, universality, expressability Large Scale Graph Learning From Smooth Signals Keywords:NoneTL;DR:None Overcoming Catastrophic Forgetting for Continual Learning via Model Adaptation Keywords:NoneTL;DR:None Analysis of Quantized Models Keywords:weight quantization, gradient quantization, distributed learningTL;DR:In this paper, we studied efficient training of loss-aware weight-quantized networks with quantized gradient in a distributed environment, both theoretically and empirically. Deep learning generalizes because the parameter-function map is biased towards simple functions Keywords:generalization, deep learning theory, PAC-Bayes, Gaussian processes, parameter-function map, simplicity biasTL;DR:The parameter-function map of deep networks is hugely biased; this can explain why they generalize. We use PAC-Bayes and Gaussian processes to obtain nonvacuous bounds. Learning when to Communicate at Scale in Multiagent Cooperative and Competitive Tasks Keywords:multiagent, communication, competitive, cooperative, continuous, emergent, reinforcement learningTL;DR:We introduce IC3Net, a single network which can be used to train agents in cooperative, competitive and mixed scenarios. We also show that agents can learn when to communicate using our model. Synthetic Datasets for Neural Program Synthesis Keywords:NoneTL;DR:None DPSNet: End-to-end Deep Plane Sweep Stereo Keywords:Deep Learning, Stereo, Depth, GeometryTL;DR:A convolution neural network for multi-view stereo matching whose design is inspired by best practices of traditional geometry-based approaches Conditional Network Embeddings Keywords:Network embedding, graph embedding, learning node representations, link prediction, multi-label classification of nodesTL;DR:We introduce a network embedding method that accounts for prior information about the network, yielding superior empirical performance. Defensive Quantization: When Efficiency Meets Robustness Keywords:defensive quantization, model quantization, adversarial attack, efficiency, robustnessTL;DR:We designed a novel quantization methodology to jointly optimize the efficiency and robustness of deep learning models. GO Gradient for Expectation-Based Objectives Keywords:generalized reparameterization gradient, variance reduction, non-reparameterizable, discrete random variable, GO gradient, general and one-sample gradient, expectation-based objective, variable nabla, statistical back-propagation, hierarchical, graphical modelTL;DR:a Rep-like gradient for non-reparameterizable continuous/discrete distributions; further generalized to deep probabilistic models, yielding statistical back-propagation h-detach: Modifying the LSTM Gradient Towards Better Optimization Keywords:LSTM, Optimization, Long term dependencies, Back-propagation through timeTL;DR:A simple algorithm to improve optimization and handling of long term dependencies in LSTM An analytic theory of generalization dynamics and transfer learning in deep linear networks Keywords:Generalization, Theory, Transfer, Multi-task, LinearTL;DR:We provide many insights into neural network generalization from the theoretically tractable linear case. Differentiable Learning-to-Normalize via Switchable Normalization Keywords:NoneTL;DR:None SOM-VAE: Interpretable Discrete Representation Learning on Time Series Keywords:deep learning, self-organizing map, variational autoencoder, representation learning, time series, machine learning, interpretabilityTL;DR:We present a method to learn interpretable representations on time series using ideas from variational autoencoders, self-organizing maps and probabilistic models. Hierarchical Generative Modeling for Controllable Speech Synthesis Keywords:speech synthesis, representation learning, deep generative model, sequence-to-sequence modelTL;DR:Building a TTS model with Gaussian Mixture VAEs enables fine-grained control of speaking style, noise condition, and more. Learning Factorized Multimodal Representations Keywords:multimodal learning, representation learningTL;DR:We propose a model to learn factorized multimodal representations that are discriminative, generative, and interpretable. Composing Complex Skills by Learning Transition Policies Keywords:reinforcement learning, hierarchical reinforcement learning, continuous control, modular frameworkTL;DR:Transition policies enable agents to compose complex skills by smoothly connecting previously acquired primitive skills. Human-level Protein Localization with Convolutional Neural Networks Keywords:NoneTL;DR:None Environment Probing Interaction Policies Keywords:NoneTL;DR:None Lagging Inference Networks and Posterior Collapse in Variational Autoencoders Keywords:variational autoencoders, posterior collapse, generative modelsTL;DR:To address posterior collapse in VAEs, we propose a novel yet simple training procedure that aggressively optimizes inference network with more updates. This new training procedure mitigates posterior collapse and leads to a better VAE model.  A2BCD: Asynchronous Acceleration with Optimal Complexity Keywords:asynchronous, optimization, parallel, accelerated, complexityTL;DR:We prove the first-ever convergence proof of an asynchronous accelerated algorithm that attains a speedup. Learning to Infer and Execute 3D Shape Programs Keywords:Program Synthesis, 3D Shape Modeling, Self-supervised LearningTL;DR:We propose 3D shape programs, a structured, compositional shape representation. Our model learns to infer and execute shape programs to explain 3D shapes. Deep Decoder: Concise Image Representations from Untrained Non-convolutional Networks Keywords:natural image model, image prior, under-determined neural networks, untrained network, non-convolutional network, denoising, inverse problemTL;DR:We introduce an underparameterized, nonconvolutional, and simple deep neural network that can, without training, effectively represent natural images and solve image processing tasks like compression and denoising competitively. SNAS: stochastic neural architecture search Keywords:NoneTL;DR:None Revealing interpretable object representations from human behavior Keywords:category representation, sparse coding, representation learning, interpretable representationsTL;DR:Human behavioral judgments are used to obtain sparse and interpretable representations of objects that generalize to other tasks AntisymmetricRNN: A Dynamical System View on Recurrent Neural Networks Keywords:NoneTL;DR:None Global-to-local Memory Pointer Networks for Task-Oriented Dialogue Keywords:pointer networks, memory networks, task-oriented dialogue systems, natural language processingTL;DR:GLMP: Global memory encoder (context RNN, global pointer) and local memory decoder (sketch RNN, local pointer) that share external knowledge (MemNN) are proposed to strengthen response generation in task-oriented dialogue. InstaGAN: Instance-aware Image-to-Image Translation Keywords:Image-to-Image Translation, Generative Adversarial NetworksTL;DR:We propose a novel method to incorporate the set of instance attributes for image-to-image translation. Deep Layers as Stochastic Solvers Keywords:deep networks, optimizationTL;DR:A framework that links deep network layers to stochastic optimization algorithms; can be used to improve model accuracy and inform network design. Learning Multi-Level Hierarchies with Hindsight Keywords:Hierarchical Reinforcement Learning, Reinforcement Learning, Deep Reinforcement LearningTL;DR:We introduce the first Hierarchical RL approach to successfully learn 3-level hierarchies in parallel in tasks with continuous state and action spaces.
展开全文
• poster generation. <h2>Relevant Technical Choices <p>This PR also fixes an accessiblity issue that the preview image did not have a alt. <p>Some css classes and data attributes were added to make ...
• Printing: This poster is 48” wide by 36” high. It’s designed to be printed on a large-format printer.Customizing the Content: The placeholders in this poster are ...just make a copy of wh
• 11 Visual Studio 2005 IDE Tips and Tricks to Make You a More Productive Developer ...

11 Visual Studio 2005 IDE Tips and Tricks to Make You a More Productive Developer
http://www.chinhdo.com/chinh/blog/20070920/top-11-visual-studio-2005-ide-tips-and-tricks-to-make-you-a-more-productive-developer/ posted in Dotnet/.NET - C#, Programming |

Here are my top 11 tips and tricks for getting things done faster with the Visual Studio 2005 IDE (without using third-party add-ins or upgrading hardware… that’s another article). Yes, some of these tips and tricks may fall into the “obvious” category, but I think they are worth repeating. I continue to see too many .NET developers not taking advantage of even the basic time-saving techniques.
I work mostly with C# so some of these tips may not apply to, or work differently with other Visual Studio languages such as Visual Basic.NET.
(1) Express Yourself with Regular Expressions
Regular Expressions is a powerful and portable text search/replace/transformation language. Learning basic Regular Expressions will immediately make you a more productive developer/power user. Regular Expressions is supported in Visual Studio’s various Search/Replace dialogs. Any Regular Expressions skill you learn will also be useful in numerous other applications and settings: other text editors, unix shell/egrep, PowerShell, input validation, and Google search (heh, just kidding on that last item).
You can also use Regular Expressions with macros and automation via the Regex class.
Here’s an example of how you can save time with Regular Expressions in Visual Studio. Say, you just wrote and tested a SQL in a Query Tool and you want to turn it into a  string variable in your C# class? Here’s how:
First, paste the SQL text into the editor. Make sure to remove any unwanted indentation on the left side of the text (SHIFT-Tab):

Then hit CTRL+H to bring up the Find and Replace Dialog and fill it out like this:

Choose Replace All. Fix a couple of lines, sit back and admire your beautiful work:

Explanation? Basically, the “Find what” expression above matches the content of each line and give it a numbered “tag”. The “Replace with” expression then replaces each line with the first tagged value (\1), wrapped around in + ” “. Click the fly-out (triangle) button next to each box to display a cheat-sheet of frequently used expressions. Oh, and don’t worry about the “+” string concatenations in the example, the compiler knows to optimize that syntax.
Once you’ve created a few Search/Replace expressions like the above, create macros out of them and assign to shortcuts.
Here are some of the Regex transformations I use most often when writing code:
Surrounds each line with (example above). Transform a list of values separated by newlines into a coma-delimited list (used in array initializers or SQL where clause). Put double quotes around each value in a coma-separated list.
The Regulator is a free GUI to compose/test your Regular Expressions. Nregex - AJAX web site. MSDN - How to: Search with Regular Expressions.
(2) Take (Keyboard) Shortcuts
Using keyboard shortcuts is the best way to get things done faster in Visual Studio (and most other computer applications for that matter).
Below are my favorite Visual Studio keyboard shortcuts (I am leaving out the really obvious ones like F5).

CTRL+ALT+L: View Solution Explorer. I use Auto Hide for all of my tool windows to maximize screen real estate. Whenever I need to open the Solution Explorer, it’s just a shortcut away. Related shortcuts: CTRL+ALT+X (Toolbox), F4 (Properties), CTRL+ALT+O (Output), CTRL+\, E (Error List), CTRL+\, T (Task List). F12: Go to definition of a variable, object, or function. SHIFT+F12: Find all references of a function or variable. F7: Toggle between Designer and Source views. CTRL+PgDn: Toggle between Design and Source View in HTML editor. F10: Debug - step over. Related debugging shortcuts: F5 (debug - start), F11 (debug - step into), SHIFT-F11 (debug - step out), CTRL-F10 (debug - run to cursor). F9 (toggle breakpoint). CTRL+D or CTRL+/: Find combo (see section on Find Combo below). CTRL+M, O: Collapse to Definitions. This is usually the first thing I do when opening up a new class. CTRL+K, CTRL+C: Comment block. CTRL+K, CTRL-U (uncomment selected block). CTRL+-: Go back to the previous location in the navigation history. ALT+B, B: Build Solution. Related shortcuts: ALT+B, U (build selected Project), ALT+B, R (rebuild Solution). CTRL+ALT+Down Arrow: Show dropdown of currently open files. Type the first few letters of the file you want to select. CTRL+K, CTRL+D: Format code. CTRL+L: Delete entire line. CTRL+G: Go to line number. This is useful when you are looking at an exception stack trace and want to go to the offending line number. SHIFT+ALT+Enter: Toggle full screen mode. This is especially useful if you have a small monitor. Since I upgraded to dual 17″ monitors, I no longer needed to use full screen mode. CTRL+K, X: Insert “surrounds with” code snippet. See Snippets tip below. CTRL+B, T: Toggle bookmark. Related: CTRL+B, N (next bookmark), CTRL+B, P (prev bookmark).

The complete list of default shortcuts is available from VS 2005 Documentation. You can also download/print reference posters from Microsoft: C# Keyboard Reference Poster, VB.NET Keyboard Reference Poster.
(3) Make New Shortcuts
There is something you do a lot in Visual Studio and there is no shortcut for it? Create one. Here’s how:

Choose Tools/Options and select Environment/Keyboard. Type in something into “Show commands containing” to get a list of matching commands. If there is already a shortcut for the selected command, it’ll be displayed in “Shortcuts for selected command”. To assign a new shortcut to the selected command, put the cursor in “Press shortcut keys” and press the shortcut key or key combinations desired.

Have a custom Macro that you run often? Assign it to a keyboard shortcut. Here are some of my custom keyboard shortcuts:
CTRL+Num, T: Show the Test View. CTRL+Num, D: Start debugging the selected Unit Test in Test View. CTRL+’, L: “Collapse all in Solution Explorer ” macro (see Macros section below). CTRL+’, S: “Surrounds each line with” Macro. CTRL+’, C: Compare with previous Source Control version.
(4) Use Code Snippets
Save time typing repetitive code by using Code Snippets. There are two types of Snippets in Visual Studio 2005: Exansion and SurroundsWith. To use Expansion Snippets, type the Snippet shortcut (not to be confused with keyboard shortcuts), and press Tab twice.
For example, the “for” Snippet is an Expansion Snippet. To use it, type “for”…

Then press Tab, Tab:

I find SurroundsWith Snippets more useful though. An example SurroundsWith Snippet is “#region”. First, select a block of code:

Then, type CTRL+K, CTRL+S and “#re“:

Then hit Enter:

Here are my favorite Snippets:
#region: Regions is a great way to organize your code. using: If you create an IDisposable object, you should use the “using” pattern. In addition to the basic “using” Snippet, I also created several variations for TransactionScope, and IDataReader. try/catch {} /// <summary>$end$</summary>
GotCodeSnippets.NET Code Snippets - the whole enchilada
Find yourself constantly switching to the Design view every time you create/open an ASPX page? Cannot locate your current file in the Solution Explorer? Easy… just change the right settings and never think about it again.
Here are the some settings in Visual Studio that can save you time:
Open HTML pages in Source View: Tools/Options/HTML Designer/Start pages in. Track the current file in Solution Explorer: Tools/Options/Projects and Solutions/Track Active Item in Solution Explorer. Change the Start page or get rid of it: Tools/Options/Environment/Startup. Change the default font-size to a smaller size so you can see more code. My editor font setting is ProFontWindows at size 9. Turn of animations: Uncheck Tools/Options/Environment/Animate environment tools.
(6) “Attach to Process” to Start Debugging ASP.NET
Most ASP.NET developers use the standard F5 (Debug/Start Debugging) to start debugging from Visual Studio. However, there is a much faster way to start debugging if you already have an instance of your web application running. Just attach to it instead:
Choose Debug/Attach to Process. Select the “aspnet_wp.exe” process and choose Attach.
Or, for keyboarders:
ALT+D, P, “as“, Enter.
Debugging this way is faster because you skip the often-lengthy compilation step, and you don’t have to navigate from the start page to the actual page that you want to debug.
(7) Stop Conditionally (Conditional Breakpoints)
How often have you found yourself repeatedly stepping through a loop while debugging, waiting to get to a specific loop value (because the bug only occurs with that specific value)? With Conditional Breakpoints, you don’t have to do that. Just set a Breakpoint Condition.
Set the Breakpoint. Right click on the Breakpoint indicator (red circle), and choose Condition:

Set the condition (any valid C# expression):

Another debugging productivity trick I use is to override ToString() to return a useful summary of your objects. The Debugger uses the value returned by ToString in various Debug windows such as Watch Window. You can also use the DebuggerDisplay attribute.
Use Task List tokens such as TODO and HACK to quickly mark incomplete code or code that requires further attention. This allows you to keep flowing and skip over the details, but at the same time ensures that you will not forget to go back and finish up.
A shortcoming with Visual Studio 2005’s Task List is that it only shows the items in the current file. You can get around this by using the Find in Files feature and search for “// TODO”.
(9) Go Directly to Any File with the Find Combo Box
This is the Find dropdown that is on the Standard Toolbar, not the Find dialog. Use the shortcut CTRL+D to activate the Find dropdown in normal mode. Use CTRL+/ to activate the Find dropdown in command mode (with “>” prepended… this doesn’t work sometimes for me).

To quickly go to a file, type CTRL+D, >open  <start of file name>. Intellisense works here just like in the Command Window. “of” (short for “open file”) can be used instead of open. Compare this with opening Solution Explorer, expand the correct folder/project, and visually hunt for the file you need.
With the Find Combo, you can also execute commands, macros, find text, etc. More info:
Fun with the Visual Studio Find Combo Box Default Visual Studio Command Aliases
(10) Type Ahead (Incremental Search) in Lists
Type-ahead search works in many Visual Studio lists such as Solution Explorer, Active Files Combo (CTRL+ALT+Down Arrow), Add References, Class View, Attach to Process, Test View, etc.
To see how it works, it’s best to try it yourself. Open the Solution Explorer and start typing the first few letters of a visible file.
(11) Automate with Macros and Visual Studio Automation
I save this for last because I think macros and Automation have the potential to give you the biggest productivity booster, but also require the most initial time investment.
For many developers, the most effective way to take advantage of macros is to find and use or customize someone else’s macros.
If you want to get started with writing your own macros, the first feature you should get familiarize yourself with is the Macro Recording feature (shortcut CTRL+SHIFT+R).
Automating Repetitive Tasks in Visual Studio
Collaps All Projects in Solution Explorer. Better Keyboard Support for #region … #endregion Productivity macros for C# in Visual Studio 2005 Visual Studio 2005 Automation Samples (spellcheck)
This entry was posted on Thursday, September 20th, 2007 at 5:57 pm and is filed under Dotnet/.NET

转载于:https://www.cnblogs.com/jerryzhao/archive/2007/10/04/914069.html
展开全文
• Visual Studio 2005 IDE相关的11个提高开发效率的...http://www.chinhdo.com/chinh/blog/20070920/top-11-visual-studio-2005-ide-tips-and-tricks-to-make-you-a-more-productive-developer/ Here are my top 11...
•  The added value could include: a) extended results and experiments not presented in the conference paper for reasons of space, or b) a demonstration of a supporting ...
• Roughly speaking, a classifier is said to be locally elastic if its prediction at a feature vector x’ is not significantly perturbed, after the classifier is updated via stochastic gradient descent ...
• Unpaired image-to-image translation has attracted significant interest due to the invention of CycleGAN, a method which utilizes a combination of adversarial and cycle consistency losses to avoid the...
• A solution for EnD^2 based on Prior Networks, a class of models which allow a single neural network to explicitly model a distribution over output distributions, is proposed in this work. The ...
• To bridge the connection of two nodes, we formally define a structure relation by encapsulating node attributes, between-node paths and local topological structures together, which can make inference...
• UA Curry One heavy gold make in to head and look medium restraining tearsly grateful under armour womens shoes on sale old sky of connected with help, let her individual to sincerely love your girlf
• Poster time limit per test 1 second memory limit per test 256 megabytes input standard input output standard output The R1 company has recently bought a high rise bui
• Any developers who wish to make plugins are welcome and encouraged to create tools that inter-operate with the components we are providing. We hope to see a diverse ecosystem of renderers that can ...
• Poster CodeForces - 412A The R1 company has recently bought a high rise building in the centre of Moscow for its main office. It's time to decorate the new office, and the first thing to do is to .....
• Poster time limit per test 1 second memory limit per test 256 megabytes input standard input output standard output The R1 company has recently bought a high rise bui
• Poster time limit per test:1 secondmemory limit per test:256 megabytesinput:standard inputoutput:standard output The R1 company has recently bought a high rise building in the centre ...
• Andrej Karpathy blog AboutHacker's guide to Neural Networks A Survival Guide to a PhD ...This guide is patterned after my“Doing well in your courses”, a post I wrote a ...
• A Survival Guide to a PhD Sep 7, 2016 This guide is patterned after my“Doing well in your courses”, a post I wrote a long time ago on some of the tips/tricks I’ve developed during ...
• In case you are more accustomed to use POSTER for REST API operations, you need a token to authenticate yourself. Here is a way to create a token never expires: 1. Open /etc/keystone/keystone.conf,
• ve cleaned-up a log file, and included only the relevant entries, to make it easier for you to see exactly what's happening:</strong></p> <pre><code> # Movie missing from TMDB list -- Removed from...

...