ICLR 2021 Schedule

Filter Events

MON 3 MAY

9 a.m.

Remarks:

Opening Remarks

(ends 10:00 AM)

10 a.m.

Poster Session 01 [10:00-12:00]

Posters 10:00-12:00

A Good Image Generator Is What You Need for High-Resolution Video Synthesis

Does enhanced shape bias improve neural network robustness to common corruptions?

Domain Generalization with MixStyle

Gauge Equivariant Mesh CNNs: Anisotropic convolutions on geometric graphs

Implicit Normalizing Flows

Isometric Transformation Invariant and Equivariant Graph Convolutional Networks

LEAF: A Learnable Frontend for Audio Classification

Learning N:M Fine-grained Structured Sparse Neural Networks From Scratch

Mutual Information State Intrinsic Control

Neural Jump Ordinary Differential Equations: Consistent Continuous-Time Prediction and Filtering

On the Universality of the Double Descent Peak in Ridgeless Regression

Progressive Skeletonization: Trimming more fat from a network at initialization

PSTNet: Point Spatio-Temporal Convolution on Point Cloud Sequences

QPLEX: Duplex Dueling Multi-Agent Q-Learning

Randomized Ensembled Double Q-Learning: Learning Fast Without a Model

Rapid Neural Architecture Search by Learning to Generate Graphs from Datasets

Rethinking the Role of Gradient-based Attribution Methods for Model Interpretability

SALD: Sign Agnostic Learning with Derivatives

SaliencyMix: A Saliency Guided Data Augmentation Strategy for Better Regularization

Towards Robust Neural Networks via Close-loop Control

WaNet - Imperceptible Warping-based Backdoor Attack

Wasserstein-2 Generative Networks

Wasserstein Embedding for Graph Learning

A Unified Approach to Interpreting and Boosting Adversarial Transferability

Batch Reinforcement Learning Through Continuation Method

Deciphering and Optimizing Multi-Task Learning: a Random Matrix Approach

Exploring Balanced Feature Spaces for Representation Learning

FairFil: Contrastive Neural Debiasing Method for Pretrained Text Encoders

Improve Object Detection with Feature-based Knowledge Distillation: Towards Accurate and Efficient Detectors

Interpreting and Boosting Dropout from a Game-Theoretic View

Is Label Smoothing Truly Incompatible with Knowledge Distillation: An Empirical Study

MELR: Meta-Learning via Modeling Episode-Level Relationships for Few-Shot Learning

Meta-GMVAE: Mixture of Gaussian VAE for Unsupervised Meta-Learning

MetaNorm: Learning to Normalize Few-Shot Batches Across Domains

MODALS: Modality-agnostic Automated Data Augmentation in the Latent Space

Neural Approximate Sufficient Statistics for Implicit Models

Noise against noise: stochastic label noise helps combat inherent label noise

On InstaHide, Phase Retrieval, and Sparse Matrix Factorization

On Learning Universal Representations Across Languages

On the Transfer of Disentangled Representations in Realistic Settings

Overfitting for Fun and Profit: Instance-Adaptive Data Compression

Parameter-Based Value Functions

Predicting Infectiousness for Proactive Contact Tracing

ResNet After All: Neural ODEs and Their Numerical Solution

Revisiting Locally Supervised Learning: an Alternative to End-to-end Training

Scalable Transfer Learning with Expert Models

Semantic Re-tuning with Contrastive Tension

Set Prediction without Imposing Structure as Conditional Density Estimation

Signatory: differentiable computations of the signature and logsignature transforms, on both CPU and GPU

Solving Compositional Reinforcement Learning Problems via Task Reduction

Spatially Structured Recurrent Modules

Stabilized Medical Image Attacks

Targeted Attack against Deep Neural Networks via Flipping Limited Weight Bits

Temporally-Extended ε-Greedy Exploration

The Unreasonable Effectiveness of Patches in Deep Convolutional Kernels Methods

Tomographic Auto-Encoder: Unsupervised Bayesian Recovery of Corrupted Data

Towards Impartial Multi-task Learning

Towards Robustness Against Natural Language Word Substitutions

Training with Quantization Noise for Extreme Model Compression

Trusted Multi-View Classification

Uncertainty Estimation and Calibration with Finite-State Probabilistic RNNs

What Makes Instance Discrimination Good for Transfer Learning?

(ends 12:00 PM)

noon

Oral Session 1 [12:00-3:05]

Ozan Sener, Yutian Chen, Blake Richards

Orals 12:00-12:30

[12:00] Dataset Condensation with Gradient Matching

[12:15] Free Lunch for Few-shot Learning: Distribution Calibration

Spotlights 12:30-12:50

[12:30] Deciphering and Optimizing Multi-Task Learning: a Random Matrix Approach

[12:40] Generalization in data-driven models of primary visual cortex

Q&As 12:50-1:00

[12:50] Q&A

Orals 1:00-1:30

[1:00] Towards Nonlinear Disentanglement in Natural Data with Temporal Sparse Coding

[1:15] A Distributional Approach to Controlled Text Generation

Spotlights 1:30-1:50

[1:30] The Intrinsic Dimension of Images and Its Impact on Learning

[1:40] How Benign is Benign Overfitting ?

Q&As 1:50-2:00

[1:50] Q&A

Orals 2:00-2:45

[2:00] Geometry-aware Instance-reweighted Adversarial Training

[2:15] Do 2D GANs Know 3D Shape? Unsupervised 3D Shape Reconstruction from 2D Image GANs

[2:30] Rethinking the Role of Gradient-based Attribution Methods for Model Interpretability

Spotlights 2:45-2:55

[2:45] Contrastive Divergence Learning is a Time Reversal Adversarial Game

Q&As 2:55-3:05

[2:55] Q&A

(ends 3:05 PM)

3:15 p.m.

Break:

BREAK:

(ends 5:00 PM)

5 p.m.

Invited Talk:

Moving beyond the fairness rhetoric in machine learning

Timnit Gebru

(ends 6:00 PM)

6 p.m.

Poster Session 02 [6:00-8:00]

Posters 6:00-8:00

Image Augmentation Is All You Need: Regularizing Deep Reinforcement Learning from Pixels

Into the Wild with AudioScope: Unsupervised Audio-Visual Separation of On-Screen Sounds

Language-Agnostic Representation Learning of Source Code from Structure and Context

Learning Structural Edits via Incremental Tree Transformations

MoVie: Revisiting Modulated Convolutions for Visual Counting and Beyond

Multi-Time Attention Networks for Irregularly Sampled Time Series

On the Stability of Fine-tuning BERT: Misconceptions, Explanations, and Strong Baselines

Overparameterisation and worst-case generalisation: friend or foe?

PAC Confidence Predictions for Deep Neural Network Classifiers

Plan-Based Relaxed Reward Shaping for Goal-Directed Tasks

Revisiting Few-sample BERT Fine-tuning

Scaling Symbolic Methods using Gradients for Neural Model Explanation

Self-Supervised Policy Adaptation during Deployment

Shape-Texture Debiased Neural Network Training

Shapley Explanation Networks

Share or Not? Learning to Schedule Language-Specific Capacity for Multilingual Translation

Single-Photon Image Classification

The Risks of Invariant Risk Minimization

The role of Disentanglement in Generalisation

The Traveling Observer Model: Multi-task Learning Through Spatial Variable Embeddings

Training GANs with Stronger Augmentations via Contrastive Discriminator

Universal approximation power of deep residual neural networks via nonlinear control theory

What Can You Learn From Your Muscles? Learning Visual Representation from Human Interactions

Accelerating Convergence of Replica Exchange Stochastic Gradient MCMC via Variance Reduction

Adaptive Federated Optimization

A statistical theory of cold posteriors in deep neural networks

Conditional Negative Sampling for Contrastive Learning of Visual Representations

Disentangling 3D Prototypical Networks for Few-Shot Concept Learning

Effective Distributed Learning with Random Features: Improved Bounds and Algorithms

Extracting Strong Policies for Robotics Tasks from Zero-Order Trajectory Optimizers

Fast convergence of stochastic subgradient method under interpolation

Gradient Projection Memory for Continual Learning

Graph Convolution with Low-rank Learnable Local Filters

GraPPa: Grammar-Augmented Pre-Training for Table Semantic Parsing

Grounding Physical Concepts of Objects and Events Through Dynamic Visual Reasoning

InfoBERT: Improving Robustness of Language Models from An Information Theoretic Perspective

Intrinsic-Extrinsic Convolution and Pooling for Learning on 3D Protein Structures

LambdaNetworks: Modeling long-range Interactions without Attention

Learning-based Support Estimation in Sublinear Time

Learning explanations that are hard to vary

Learning from others' mistakes: Avoiding dataset biases without modeling them

Learning Hyperbolic Representations of Topological Features

Learning Invariant Representations for Reinforcement Learning without Reconstruction

Learning "What-if" Explanations for Sequential Decision-Making

Learning with AMIGo: Adversarially Motivated Intrinsic Goals

LiftPool: Bidirectional ConvNet Pooling

Multi-Level Local SGD: Distributed SGD for Heterogeneous Hierarchical Networks

MultiModalQA: complex question answering over text, tables and images

NeMo: Neural Mesh Models of Contrastive Features for Robust 3D Pose Estimation

Off-Dynamics Reinforcement Learning: Training for Transfer with Domain Classifiers

On Statistical Bias In Active Learning: How and When to Fix It

On the Impossibility of Global Convergence in Multi-Loss Optimization

On the role of planning in model-based deep reinforcement learning

On the Universality of Rotation Equivariant Point Cloud Networks

Open Question Answering over Tables and Text

Parameter Efficient Multimodal Transformers for Video Representation Learning

Planning from Pixels using Inverse Dynamics Models

Predicting Classification Accuracy When Adding New Unobserved Classes

Predicting Inductive Biases of Pre-Trained Models

Primal Wasserstein Imitation Learning

Pruning Neural Networks at Initialization: Why Are We Missing the Mark?

Rapid Task-Solving in Novel Environments

Representation learning for improved interpretability and classification accuracy of clinical factors from EEG

Reset-Free Lifelong Learning with Skill-Space Planning

Rethinking Embedding Coupling in Pre-trained Language Models

Saliency is a Possible Red Herring When Diagnosing Poor Generalization

Selectivity considered harmful: evaluating the causal impact of class selectivity in DNNs

Seq2Tens: An Efficient Representation of Sequences by Low-Rank Tensor Projections

Shapley explainability on the data manifold

Sparse encoding for more-interpretable feature-selecting representations in probabilistic matrix factorization

Structured Prediction as Translation between Augmented Natural Languages

Symmetry-Aware Actor-Critic for 3D Molecular Design

Teaching Temporal Logics to Neural Networks

Towards Nonlinear Disentanglement in Natural Data with Temporal Sparse Coding

Trajectory Prediction using Equivariant Continuous Convolution

Uncertainty Sets for Image Classifiers using Conformal Prediction

Understanding the failure modes of out-of-distribution generalization

Unsupervised Meta-Learning through Latent-Space Interpolation in Generative Models

Using latent space regression to analyze and leverage compositionality in GANs

Vector-output ReLU Neural Network Problems are Copositive Programs: Convex Analysis of Two Layer Networks and Polynomial-time Algorithms

What Should Not Be Contrastive in Contrastive Learning

WrapNet: Neural Net Inference with Ultra-Low-Precision Arithmetic

X2T: Training an X-to-Text Typing Interface with Online Learning from User Feedback

Zero-Cost Proxies for Lightweight NAS

(ends 8:00 PM)

8 p.m.

Oral Session 2 [8:00-11:23]

Orals 8:00-8:45

[8:00] Federated Learning Based on Dynamic Regularization

[8:15] Gradient Projection Memory for Continual Learning

[8:30] Growing Efficient Deep Networks by Structured Continuous Sparsification

Spotlights 8:45-8:55

[8:45] Geometry-Aware Gradient Algorithms for Neural Architecture Search

Q&As 8:55-9:05

[8:55] Q&A

Spotlights 9:05-10:05

[9:05] Generalization bounds via distillation

[9:15] On the Theory of Implicit Deep Learning: Global Convergence with Implicit Layers

[9:25] Sharpness-aware Minimization for Efficiently Improving Generalization

[9:35] Systematic generalisation with group invariant predictions

[9:45] On Statistical Bias In Active Learning: How and When to Fix It

[9:55] Very Deep VAEs Generalize Autoregressive Models and Can Outperform Them on Images

Q&As 10:05-10:20

[10:05] Q&A

Spotlights 10:20-11:10

[10:20] Uncertainty Sets for Image Classifiers using Conformal Prediction

[10:30] PMI-Masking: Principled masking of correlated spans

[10:40] Gradient Vaccine: Investigating and Improving Multi-task Optimization in Massively Multilingual Models

[10:50] Watch-And-Help: A Challenge for Social Perception and Human-AI Collaboration

[11:00] Predicting Infectiousness for Proactive Contact Tracing

Q&As 11:10-11:23

[11:10] Q&A

(ends 11:23 PM)

11:30 p.m.

Break:

BREAK:

(ends 1:00 AM)

TUE 4 MAY

1 a.m.

Invited Talk:

Commonsense AI: Myth and Truth

Yejin Choi

(ends 2:00 AM)

2 a.m.

Poster Session 03 [2:00-4:00]

Posters 2:00-4:00

Benefit of deep learning with non-convex noisy gradient descent: Provable excess risk bound and superiority to kernel methods

Deberta: Decoding-Enhanced Bert With Disentangled Attention

Decentralized Attribution of Generative Models

Latent Skill Planning for Exploration and Transfer

Learning Energy-Based Models by Diffusion Recovery Likelihood

MONGOOSE: A Learnable LSH Framework for Efficient Neural Network Training

Random Feature Attention

Regularized Inverse Reinforcement Learning

Remembering for the Right Reasons: Explanations Reduce Catastrophic Forgetting

Representation Learning for Sequence Data with Deep Autoencoding Predictive Components

Robust Curriculum Learning: from clean label detection to noisy label self-correction

Sequential Density Ratio Estimation for Simultaneous Optimization of Speed and Accuracy

SOLAR: Sparse Orthogonal Learned and Random Embeddings

The Intrinsic Dimension of Images and Its Impact on Learning

When does preconditioning help or hurt generalization?

When Optimizing $f$-Divergence is Robust with Label Noise

Zero-shot Synthesis with Group-Supervised Learning

Bypassing the Ambient Dimension: Private SGD with Gradient Subspace Identification

Contextual Transformation Networks for Online Continual Learning

Deep Partition Aggregation: Provable Defenses against General Poisoning Attacks

DeLighT: Deep and Light-weight Transformer

Explaining the Efficacy of Counterfactually Augmented Data

Federated Learning Based on Dynamic Regularization

Improved Estimation of Concentration Under $\ell_p$-Norm Distance Metrics Using Half Spaces

Incorporating Symmetry into Deep Dynamics Models for Improved Generalization

Layer-adaptive Sparsity for the Magnitude-based Pruning

Learning a Latent Simplex in Input Sparsity Time

Learning A Minimax Optimizer: A Pilot Study

Long Live the Lottery: The Existence of Winning Tickets in Lifelong Learning

MALI: A memory efficient and reverse accurate integrator for Neural ODEs

Meta-Learning with Neural Tangent Kernels

MixKD: Towards Efficient Distillation of Large-scale Language Models

Model Patching: Closing the Subgroup Performance Gap with Data Augmentation

MoPro: Webly Supervised Learning with Momentum Prototypes

Neural Attention Distillation: Erasing Backdoor Triggers from Deep Neural Networks

Offline Model-Based Optimization via Normalized Maximum Likelihood Estimation

On Dyadic Fairness: Exploring and Mitigating Bias in Graph Connections

One Network Fits All? Modular versus Monolithic Task Formulations in Neural Networks

On Fast Adversarial Robustness Adaptation in Model-Agnostic Meta-Learning

Online Adversarial Purification based on Self-supervised Learning

On the geometry of generalization and memorization in deep neural networks

Optimal Regularization can Mitigate Double Descent

Optimizing Memory Placement using Evolutionary Graph Reinforcement Learning

Parrot: Data-Driven Behavioral Priors for Reinforcement Learning

Partitioned Learned Bloom Filters

Personalized Federated Learning with First Order Model Optimization

PlasticineLab: A Soft-Body Manipulation Benchmark with Differentiable Physics

Proximal Gradient Descent-Ascent: Variable Convergence under KŁ Geometry

PseudoSeg: Designing Pseudo Labels for Semantic Segmentation

Regularization Matters in Policy Optimization - An Empirical Study on Continuous Control

Rethinking Architecture Selection in Differentiable NAS

Rethinking Positional Encoding in Language Pre-training

Robust and Generalizable Visual Representation Learning via Random Convolutions

Robust Reinforcement Learning on State Observations with Learned Optimal Adversary

SAFENet: A Secure, Accurate and Fast Neural Network Inference

Score-Based Generative Modeling through Stochastic Differential Equations

SCoRe: Pre-Training for Context Representation in Conversational Semantic Parsing

Selective Classification Can Magnify Disparities Across Groups

Self-training For Few-shot Transfer Across Extreme Task Differences

Semi-supervised Keypoint Localization

SenSeI: Sensitive Set Invariance for Enforcing Individual Fairness

Spatio-Temporal Graph Scattering Transform

Taking Notes on the Fly Helps Language Pre-Training

The Deep Bootstrap Framework: Good Online Learners are Good Offline Generalizers

The Role of Momentum Parameters in the Optimal Convergence of Adaptive Polyak's Heavy-ball Methods

Tilted Empirical Risk Minimization

Undistillable: Making A Nasty Teacher That CANNOT teach students

Unlearnable Examples: Making Personal Data Unexploitable

UPDeT: Universal Multi-agent RL via Policy Decoupling with Transformers

VA-RED$^2$: Video Adaptive Redundancy Reduction

Variational Intrinsic Control Revisited

Very Deep VAEs Generalize Autoregressive Models and Can Outperform Them on Images

WaveGrad: Estimating Gradients for Waveform Generation

What are the Statistical Limits of Offline RL with Linear Function Approximation?

Why resampling outperforms reweighting for correcting sampling bias with stochastic gradients

(ends 4:00 AM)

4 a.m.

Oral Session 3 [4:00-7:16]

Orals 4:00-4:45

[4:00] SMiRL: Surprise Minimizing Reinforcement Learning in Unstable Environments

[4:15] Contrastive Explanations for Reinforcement Learning via Embedded Self Predictions

[4:30] Parrot: Data-Driven Behavioral Priors for Reinforcement Learning

Spotlights 4:45-5:05

[4:45] Structured Prediction as Translation between Augmented Natural Languages

[4:55] Mathematical Reasoning via Self-supervised Skip-tree Training

Q&As 5:05-5:18

[5:05] Q&A

Spotlights 5:18-6:08

[5:18] Improving Adversarial Robustness via Channel-wise Activation Suppressing

[5:28] Fast Geometric Projections for Local Robustness Certification

[5:38] Information Laundering for Model Privacy

[5:48] Dataset Inference: Ownership Resolution in Machine Learning

[5:58] HW-NAS-Bench: Hardware-Aware Neural Architecture Search Benchmark

Q&As 6:08-6:21

[6:08] Q&A

Orals 6:21-6:36

[6:21] How Neural Networks Extrapolate: From Feedforward to Graph Neural Networks

Spotlights 6:36-7:06

[6:36] Graph Convolution with Low-rank Learnable Local Filters

[6:46] The Traveling Observer Model: Multi-task Learning Through Spatial Variable Embeddings

[6:56] Meta-GMVAE: Mixture of Gaussian VAE for Unsupervised Meta-Learning

Q&As 7:06-7:16

[7:06] Q&A

(ends 7:16 AM)

7:30 a.m.

Break:

BREAK:

(ends 7:30 AM)

9 a.m.

Invited Talk:

Geometric Deep Learning: the Erlangen Programme of ML

Michael Bronstein

(ends 10:00 AM)

10 a.m.

Poster Session 04 [10:00-12:00]

Posters 10:00-12:00

A Distributional Approach to Controlled Text Generation

Bayesian Context Aggregation for Neural Processes

Capturing Label Characteristics in VAEs

Class Normalization for (Continual)? Generalized Zero-Shot Learning

Complex Query Answering with Neural Link Predictors

Contemplating Real-World Object Classification

Disambiguating Symbolic Expressions in Informal Documents

Do 2D GANs Know 3D Shape? Unsupervised 3D Shape Reconstruction from 2D Image GANs

Effective Abstract Reasoning with Dual-Contrast Network

Efficient Certified Defenses Against Patch Attacks on Image Classifiers

Exemplary Natural Images Explain CNN Activations Better than State-of-the-Art Feature Visualization

Generalization in data-driven models of primary visual cortex

Generalized Multimodal ELBO

Group Equivariant Stand-Alone Self-Attention For Vision

Identifying nonlinear dynamical systems with multiple time scales and long-range dependencies

Large-width functional asymptotics for deep Gaussian neural networks

Lossless Compression of Structured Convolutional Models via Lifting

MiCE: Mixture of Contrastive Experts for Unsupervised Image Clustering

PDE-Driven Spatiotemporal Disentanglement

Policy-Driven Attack: Learning to Query for Hard-label Black-box Adversarial Examples

Probing BERT in Hyperbolic Spaces

Sample-Efficient Automated Deep Reinforcement Learning

Scaling the Convex Barrier with Active Sets

SkipW: Resource Adaptable RNN with Strict Upper Computational Limit

A Block Minifloat Representation for Training Deep Neural Networks

Accurate Learning of Graph Representations with Graph Multiset Pooling

Activation-level uncertainty in deep neural networks

AdaGCN: Adaboosting Graph Convolutional Networks into Deep Models

A Trainable Optimal Transport Embedding for Feature Aggregation and its Relationship to Attention

A Universal Representation Transformer Layer for Few-Shot Image Classification

Auxiliary Learning by Implicit Differentiation

BOIL: Towards Representation Change for Few-shot Learning

Calibration tests beyond classification

Computational Separation Between Convolutional and Fully-Connected Networks

Conformation-Guided Molecular Representation with Hamiltonian Neural Networks

Coping with Label Shift via Distributionally Robust Optimisation

Deep Repulsive Clustering of Ordered Data Based on Order-Identity Decomposition

Distributed Momentum for Byzantine-resilient Stochastic Gradient Descent

FedMix: Approximation of Mixup under Mean Augmented Federated Learning

Generalized Energy Based Models

GraphCodeBERT: Pre-training Code Representations with Data Flow

Group Equivariant Conditional Neural Processes

Hyperbolic Neural Networks++

IDF++: Analyzing and Improving Integer Discrete Flows for Lossless Compression

Image GANs meet Differentiable Rendering for Inverse Graphics and Interpretable 3D Neural Rendering

Improving Transformation Invariance in Contrastive Representation Learning

Intraclass clustering: an implicit learning ability that regularizes DNNs

Learning Accurate Entropy Model with Global Reference for Image Compression

Learning Better Structured Representations Using Low-rank Adaptive Label Smoothing

Learning Incompressible Fluid Dynamics from Scratch - Towards Fast, Differentiable Fluid Models that Generalize

Learning Subgoal Representations with Slow Dynamics

Learning the Pareto Front with Hypernetworks

Model-based micro-data reinforcement learning: what are the crucial model properties and which model to choose?

Monte-Carlo Planning and Learning with Language Action Value Estimates

Multiscale Score Matching for Out-of-Distribution Detection

Neurally Augmented ALISTA

not-MIWAE: Deep Generative Modelling with Missing not at Random Data

On Self-Supervised Image Representations for GAN Evaluation

Practical Real Time Recurrent Learning with a Sparse Approximation

Prediction and generalisation over directed actions by grid cells

Prototypical Contrastive Learning of Unsupervised Representations

Rao-Blackwellizing the Straight-Through Gumbel-Softmax Gradient Estimator

Refining Deep Generative Models via Discriminator Gradient Flow

Risk-Averse Offline Reinforcement Learning

Spatial Dependency Networks: Neural Layers for Improved Generative Image Modeling

What they do when in doubt: a study of inductive biases in seq2seq learners

Winning the L2RPN Challenge: Power Grid Management via Semi-Markov Afterstate Actor-Critic

(ends 12:00 PM)

noon

Oral Session 4 [12:00-2:58]

Orals 12:00-12:15

[12:00] End-to-end Adversarial Text-to-Speech

Spotlights 12:15-12:55

[12:15] Autoregressive Entity Retrieval

[12:25] Interpreting Graph Neural Networks for NLP With Differentiable Edge Masking

[12:35] Expressive Power of Invariant and Equivariant Graph Neural Networks

[12:45] Gauge Equivariant Mesh CNNs: Anisotropic convolutions on geometric graphs

Q&As 12:55-1:08

[12:55] Q&A

Orals 1:08-1:38

[1:08] Rao-Blackwellizing the Straight-Through Gumbel-Softmax Gradient Estimator

[1:23] Scalable Learning and MAP Inference for Nonsymmetric Determinantal Point Processes

Spotlights 1:38-1:58

[1:38] Multivariate Probabilistic Time Series Forecasting via Conditioned Normalizing Flows

[1:48] Noise against noise: stochastic label noise helps combat inherent label noise

Q&As 1:58-2:08

[1:58] Q&A

Spotlights 2:08-2:48

[2:08] Mutual Information State Intrinsic Control

[2:18] Learning Incompressible Fluid Dynamics from Scratch - Towards Fast, Differentiable Fluid Models that Generalize

[2:28] Identifying nonlinear dynamical systems with multiple time scales and long-range dependencies

[2:38] Fidelity-based Deep Adiabatic Scheduling

Q&As 2:48-2:58

[2:48] Q&A

(ends 2:58 PM)

3 p.m.

Break:

BREAK:

(ends 5:00 PM)

5 p.m.

Invited Talk:

AI in Finance: Scope and Examples

Manuela Veloso

(ends 6:00 PM)

6 p.m.

Poster Session 05 [6:00-8:00]

Posters 6:00-8:00

Are wider nets better given the same number of parameters?

DC3: A learning method for optimization with hard constraints

DINO: A Conditional Energy-Based GAN for Domain Translation

Direction Matters: On the Implicit Bias of Stochastic Gradient Descent with Moderate Learning Rate

Large Associative Memory Problem in Neurobiology and Machine Learning

Quantifying Differences in Reward Functions

Rank the Episodes: A Simple Approach for Exploration in Procedurally-Generated Environments

Recurrent Independent Mechanisms

Reducing the Computational Cost of Deep Generative Models with Binary Neural Networks

Representation Learning via Invariant Causal Mechanisms

Robust Pruning at Initialization

Shape or Texture: Understanding Discriminative Features in CNNs

Sharper Generalization Bounds for Learning with Gradient-dominated Objective Functions

Systematic generalisation with group invariant predictions

Tent: Fully Test-Time Adaptation by Entropy Minimization

Text Generation by Learning from Demonstrations

The geometry of integration in text classification RNNs

Towards Resolving the Implicit Bias of Gradient Descent for Matrix Factorization: Greedy Low-Rank Learning

Unsupervised Object Keypoint Learning using Local Spatial Predictability

VAEBM: A Symbiosis between Variational Autoencoders and Energy-based Models

Anatomy of Catastrophic Forgetting: Hidden Representations and Task Semantics

Approximate Nearest Neighbor Negative Contrastive Learning for Dense Text Retrieval

Auction Learning as a Two-Player Game

Characterizing signal propagation to close the performance gap in unnormalized ResNets

Clairvoyance: A Pipeline Toolkit for Medical Time Series

C-Learning: Horizon-Aware Cumulative Accessibility Estimation

Data-Efficient Reinforcement Learning with Self-Predictive Representations

Decoupling Global and Local Representations via Invertible Generative Flows

DialoGraph: Incorporating Interpretable Strategy-Graph Networks into Negotiation Dialogues

Discovering a set of policies for the worst case reward

Distance-Based Regularisation of Deep Networks for Fine-Tuning

FairBatch: Batch Selection for Model Fairness

Fair Mixup: Fairness via Interpolation

Getting a CLUE: A Method for Explaining Uncertainty Estimates

Global optimality of softmax policy gradient with single hidden layer neural networks in the mean-field regime

How Benign is Benign Overfitting ?

Interpreting Knowledge Graph Relation Representation from Word Embeddings

Iterative Empirical Game Solving via Single Policy Best Response

Learning a Latent Search Space for Routing Problems using Variational Autoencoders

Learning from Protein Structure with Geometric Vector Perceptrons

Learning Neural Event Functions for Ordinary Differential Equations

Learning Parametrised Graph Shift Operators

Learning Robust State Abstractions for Hidden-Parameter Block MDPs

Learning Value Functions in Deep Policy Gradients using Residual Variance

Mapping the Timescale Organization of Neural Language Models

Meta-learning Symmetries by Reparameterization

Nearest Neighbor Machine Translation

NOVAS: Non-convex Optimization via Adaptive Stochastic Search for End-to-end Learning and Control

On the Dynamics of Training Attention Models

On the mapping between Hopfield networks and Restricted Boltzmann Machines

On the Origin of Implicit Regularization in Stochastic Gradient Descent

On the Theory of Implicit Deep Learning: Global Convergence with Implicit Layers

Physics-aware, probabilistic model order reduction with guaranteed stability

Provable Rich Observation Reinforcement Learning with Combinatorial Latent States

Reinforcement Learning with Random Delays

Rethinking Attention with Performers

Ringing ReLUs: Harmonic Distortion Analysis of Nonlinear Feedforward Networks

Scalable Bayesian Inverse Reinforcement Learning

Self-Supervised Learning of Compressed Video Representations

Single-Timescale Actor-Critic Provably Finds Globally Optimal Policy

SMiRL: Surprise Minimizing Reinforcement Learning in Unstable Environments

SSD: A Unified Framework for Self-Supervised Outlier Detection

Statistical inference for individual fairness

Supervised Contrastive Learning for Pre-trained Language Model Fine-tuning

Support-set bottlenecks for video-text representation learning

Taming GANs with Lookahead-Minmax

Teaching with Commentaries

Towards Faster and Stabilized GAN Training for High-fidelity Few-shot Image Synthesis

Tradeoffs in Data Augmentation: An Empirical Study

Training BatchNorm and Only BatchNorm: On the Expressive Power of Random Features in CNNs

Transformer protein language models are unsupervised structure learners

Transient Non-stationarity and Generalisation in Deep Reinforcement Learning

UMEC: Unified model and embedding compression for efficient recommendation systems

Uncertainty-aware Active Learning for Optimal Bayesian Classifier

Uncertainty Estimation in Autoregressive Structured Prediction

Understanding Over-parameterization in Generative Adversarial Networks

Unsupervised Representation Learning for Time Series with Temporal Neighborhood Coding

VTNet: Visual Transformer Network for Object Goal Navigation

Vulnerability-Aware Poisoning Mechanism for Online RL with Unknown Dynamics

(ends 8:00 PM)

8 p.m.

Oral Session 5 [8:00-10:56]

Orals 8:00-8:30

[8:00] Iterated learning for emergent systematicity in VQA

[8:15] Learning Generalizable Visual Representations via Interactive Gameplay

Spotlights 8:30-8:50

[8:30] How Does Mixup Help With Robustness and Generalization?

[8:40] Recurrent Independent Mechanisms

Q&As 8:50-9:00

[8:50] Q&A

Orals 9:00-9:30

[9:00] Randomized Automatic Differentiation

[9:15] Image GANs meet Differentiable Rendering for Inverse Graphics and Interpretable 3D Neural Rendering

Spotlights 9:30-10:00

[9:30] Mind the Pad -- CNNs Can Develop Blind Spots

[9:40] Implicit Convex Regularizers of CNN Architectures: Convex Optimization of Two- and Three-Layer Networks in Polynomial Time

[9:50] Learning from Protein Structure with Geometric Vector Perceptrons

Q&As 10:00-10:13

[10:00] Q&A

Orals 10:13-10:28

[10:13] On the mapping between Hopfield networks and Restricted Boltzmann Machines

Spotlights 10:28-10:48

[10:28] Learning-based Support Estimation in Sublinear Time

[10:38] Long-tail learning via logit adjustment

Q&As 10:48-10:56

[10:48] Q&A

(ends 10:56 PM)

11 p.m.

Expo Talk Panel:

Interpretability with skeptical and user-centric mind

(ends 12:00 AM)

WED 5 MAY

midnight

Break:

BREAK:

(ends 1:00 AM)

1 a.m.

Town Hall:

Town Hall

(ends 2:00 AM)

2 a.m.

Poster Session 06 [2:00-4:00]

Posters 2:00-4:00

A Discriminative Gaussian Mixture Model with Sparsity

Autoregressive Dynamics Models for Offline Policy Evaluation and Optimization

Can a Fruit Fly Learn Word Embeddings?

Debiasing Concept-based Explanations with Causal Analysis

Deep Equals Shallow for ReLU Networks in Kernel Regimes

Diverse Video Generation using a Gaussian Process Trigger

DrNAS: Dirichlet Neural Architecture Search

Drop-Bottleneck: Learning Discrete Compressed Representation for Noise-Robust Exploration

Fourier Neural Operator for Parametric Partial Differential Equations

Fuzzy Tiling Activations: A Simple Approach to Learning Sparse Representations Online

Hierarchical Autoregressive Modeling for Neural Video Compression

Implicit Convex Regularizers of CNN Architectures: Convex Optimization of Two- and Three-Layer Networks in Polynomial Time

Lipschitz Recurrent Neural Networks

Mirostat: A Neural Text Decoding Algorithm That Directly Controls Perplexity

On Graph Neural Networks versus Graph-Augmented MLPs

SEDONA: Search for Decoupled Neural Networks toward Greedy Block-wise Learning

The Importance of Pessimism in Fixed-Dataset Policy Optimization

Topology-Aware Segmentation Using Discrete Morse Theory

Universal Weakly Supervised Segmentation by Pixel-to-Segment Contrastive Learning

Usable Information and Evolution of Optimal Representations During Training

Achieving Linear Speedup with Partial Worker Participation in Non-IID Federated Learning

A Hypergradient Approach to Robust Regression without Correspondence

Aligning AI With Shared Human Values

Are Neural Rankers still Outperformed by Gradient Boosted Decision Trees?

A Temporal Kernel Approach for Deep Learning with Continuous-time Information

Attentional Constellation Nets for Few-Shot Learning

A unifying view on implicit bias in training linear neural networks

Behavioral Cloning from Noisy Demonstrations

Bowtie Networks: Generative Modeling for Joint Few-Shot Recognition and Novel-View Synthesis

BUSTLE: Bottom-Up Program Synthesis Through Learning-Guided Exploration

CcGAN: Continuous Conditional Generative Adversarial Networks for Image Generation

CoDA: Contrast-enhanced and Diversity-promoting Data Augmentation for Natural Language Understanding

Co-Mixup: Saliency Guided Joint Mixup with Supermodular Diversity

CompOFA – Compound Once-For-All Networks for Faster Multi-Platform Deployment

Concept Learners for Few-Shot Learning

Contextual Dropout: An Efficient Sample-Dependent Dropout Module

Convex Potential Flows: Universal Probability Distributions with Optimal Transport and Convex Optimization

CopulaGNN: Towards Integrating Representational and Correlational Roles of Graphs in Graph Neural Networks

Dataset Inference: Ownership Resolution in Machine Learning

DDPNOpt: Differential Dynamic Programming Neural Optimizer

Denoising Diffusion Implicit Models

Discovering Non-monotonic Autoregressive Orderings with Variational Inference

Discrete Graph Structure Learning for Forecasting Multiple Time Series

DOP: Off-Policy Multi-Agent Decomposed Policy Gradients

Do Wide and Deep Networks Learn the Same Things? Uncovering How Neural Network Representations Vary with Width and Depth

Enjoy Your Editing: Controllable GANs for Image Editing via Latent Space Navigation

Fantastic Four: Differentiable and Efficient Bounds on Singular Values of Convolution Layers

FedBE: Making Bayesian Model Ensemble Applicable to Federated Learning

Federated Semi-Supervised Learning with Inter-Client Consistency & Disjoint Learning

Generalized Variational Continual Learning

Generating Adversarial Computer Programs using Optimized Obfuscations

Gradient Descent on Neural Networks Typically Occurs at the Edge of Stability

HalentNet: Multimodal Trajectory Forecasting with Hallucinative Intents

Hopper: Multi-hop Transformer for Spatiotemporal Reasoning

How Neural Networks Extrapolate: From Feedforward to Graph Neural Networks

How to Find Your Friendly Neighborhood: Graph Attention Design with Self-Supervision

Implicit Under-Parameterization Inhibits Data-Efficient Deep Reinforcement Learning

Individually Fair Rankings

Information Laundering for Model Privacy

Knowledge Distillation as Semiparametric Inference

Large Batch Simulation for Deep Reinforcement Learning

Learning Safe Multi-agent Control with Decentralized Neural Barrier Certificates

Learning to Reach Goals via Iterated Supervised Learning

Linear Mode Connectivity in Multitask and Continual Learning

Mathematical Reasoning via Self-supervised Skip-tree Training

Memory Optimization for Deep Networks

Monotonic Kronecker-Factored Lattice

Multi-resolution modeling of a discrete stochastic process identifies causes of cancer

Neural Mechanics: Symmetry and Broken Conservation Laws in Deep Learning Dynamics

Projected Latent Markov Chain Monte Carlo: Conditional Sampling of Normalizing Flows

RMSprop converges with proper hyper-parameter

RNNLogic: Learning Logic Rules for Reasoning on Knowledge Graphs

Understanding the role of importance weighting for deep learning

Unsupervised Discovery of 3D Physical Objects

Viewmaker Networks: Learning Views for Unsupervised Representation Learning

Why Are Convolutional Nets More Sample-Efficient than Fully-Connected Nets?

(ends 4:00 AM)

4 a.m.

Oral Session 6 [4:00-7:16]

Orals 4:00-4:15

[4:00] Deep symbolic regression: Recovering mathematical expressions from data via risk-seeking policy gradients

Spotlights 4:15-4:45

[4:15] DDPNOpt: Differential Dynamic Programming Neural Optimizer

[4:25] Orthogonalizing Convolutional Layers with the Cayley Transform

[4:35] Model-Based Visual Planning with Self-Supervised Functional Distances

Q&As 4:45-4:55

[4:45] Q&A

Orals 4:55-5:10

[4:55] Global Convergence of Three-layer Neural Networks in the Mean Field Regime

Spotlights 5:10-5:50

[5:10] Minimum Width for Universal Approximation

[5:20] Async-RED: A Provably Convergent Asynchronous Block Parallel Stochastic Method using Deep Denoising Priors

[5:30] Individually Fair Gradient Boosting

[5:40] Are Neural Rankers still Outperformed by Gradient Boosted Decision Trees?

Q&As 5:50-6:03

[5:50] Q&A

Orals 6:03-6:33

[6:03] Co-Mixup: Saliency Guided Joint Mixup with Supermodular Diversity

[6:18] MONGOOSE: A Learnable LSH Framework for Efficient Neural Network Training

Spotlights 6:33-7:03

[6:33] Locally Free Weight Sharing for Network Width Search

[6:43] Memory Optimization for Deep Networks

[6:53] Neural Topic Model via Optimal Transport

Q&As 7:03-7:16

[7:03] Q&A

(ends 7:16 AM)

7:30 a.m.

Break:

BREAK:

(ends 7:30 AM)

9 a.m.

Invited Talk:

Perceiving the 3D World from Images and Video

Lourdes Agapito

(ends 10:00 AM)

10 a.m.

Poster Session 07 [10:00-12:00]

Posters 10:00-12:00

A Panda? No, It's a Sloth: Slowdown Attacks on Adaptive Multi-Exit Neural Network Inference

Bag of Tricks for Adversarial Training

Dance Revolution: Long-Term Dance Generation with Music via Curriculum Learning

Deployment-Efficient Reinforcement Learning via Model-Based Offline Optimization

Differentiable Segmentation of Sequences

Efficient Continual Learning with Modular Networks and Task-Driven Priors

Expressive Power of Invariant and Equivariant Graph Neural Networks

Graph Edit Networks

Improving Relational Regularized Autoencoders with Spherical Sliced Fused Gromov Wasserstein

Interpretable Neural Architecture Search via Bayesian Optimisation with Weisfeiler-Lehman Kernels

Negative Data Augmentation

Neural Delay Differential Equations

Neural ODE Processes

Removing Undesirable Feature Contributions Using Out-of-Distribution Data

Return-Based Contrastive Representation Learning for Reinforcement Learning

Robust Learning of Fixed-Structure Bayesian Networks in Nearly-Linear Time

A Better Alternative to Error Feedback for Communication-Efficient Distributed Learning

Acting in Delayed Environments with Non-Stationary Markov Policies

Active Contrastive Learning of Audio-Visual Video Representations

An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale

Augmenting Physical Models with Deep Networks for Complex Dynamics Forecasting

Auxiliary Task Update Decomposition: The Good, the Bad and the Neutral

A Wigner-Eckart Theorem for Group Equivariant Convolution Kernels

Bidirectional Variational Inference for Non-Autoregressive Text-to-Speech

BRECQ: Pushing the Limit of Post-Training Quantization by Block Reconstruction

Byzantine-Resilient Non-Convex Stochastic Gradient Descent

CoCon: A Self-Supervised Approach for Controlled Text Generation

Communication in Multi-Agent Reinforcement Learning: Intention Sharing

Deep Learning meets Projective Clustering

Deep Neural Network Fingerprinting by Conferrable Adversarial Examples

Degree-Quant: Quantization-Aware Training for Graph Neural Networks

DICE: Diversity in Deep Ensembles via Conditional Redundancy Adversarial Estimation

Discovering Diverse Multi-Agent Strategic Behavior via Reward Randomization

Disentangled Recurrent Wasserstein Autoencoder

Explainable Deep One-Class Classification

Explaining by Imitating: Understanding Decisions by Interpretable Policy Learning

Fidelity-based Deep Adiabatic Scheduling

FOCAL: Efficient Fully-Offline Meta-Reinforcement Learning via Distance Metric Learning and Behavior Regularization

Fooling a Complete Neural Network Verifier

Geometry-aware Instance-reweighted Adversarial Training

Gradient Origin Networks

Grounded Language Learning Fast and Slow

High-Capacity Expert Binary Networks

Identifying Physical Law of Hamiltonian Systems via Meta-Learning

IEPT: Instance-Level and Episode-Level Pretext Tasks for Few-Shot Learning

Improving Adversarial Robustness via Channel-wise Activation Suppressing

Isometric Propagation Network for Generalized Zero-shot Learning

Knowledge distillation via softmax regression representation learning

Learning Associative Inference Using Fast Weight Memory

Learning from Demonstration with Weakly Supervised Disentanglement

Learning Task Decomposition with Ordered Memory Policy Network

Locally Free Weight Sharing for Network Width Search

Long Range Arena : A Benchmark for Efficient Transformers

Net-DNF: Effective Deep Modeling of Tabular Data

Neural gradients are near-lognormal: improved quantized and sparse training

Neural networks with late-phase weights

New Bounds For Distributed Mean Estimation and Variance Reduction

No Cost Likelihood Manipulation at Test Time for Making Better Mistakes in Deep Networks

On Data-Augmentation and Consistency-Based Semi-Supervised Learning

Relating by Contrasting: A Data-efficient Framework for Multimodal Generative Models

Reweighting Augmented Samples by Minimizing the Maximal Expected Loss

Self-supervised Adversarial Robustness for the Low-label, High-data Regime

Separation and Concentration in Deep Networks

Simple Spectral Graph Convolution

(ends 12:00 PM)

noon

Oral Session 7 [12:00-3:08]

Orals 12:00-12:45

[12:00] An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale

[12:15] Rethinking Attention with Performers

[12:30] Share or Not? Learning to Schedule Language-Specific Capacity for Multilingual Translation

Spotlights 12:45-12:55

[12:45] Support-set bottlenecks for video-text representation learning

Q&As 12:55-1:05

[12:55] Q&A

Orals 1:05-1:20

[1:05] Getting a CLUE: A Method for Explaining Uncertainty Estimates

Spotlights 1:20-1:50

[1:20] Influence Estimation for Generative Adversarial Networks

[1:30] Stabilized Medical Image Attacks

[1:40] Deep Neural Network Fingerprinting by Conferrable Adversarial Examples

Q&As 1:50-2:00

[1:50] Q&A

Orals 2:00-2:15

[2:00] Learning Cross-Domain Correspondence for Control with Dynamics Cycle-Consistency

Spotlights 2:15-2:55

[2:15] Benefit of deep learning with non-convex noisy gradient descent: Provable excess risk bound and superiority to kernel methods

[2:25] Tent: Fully Test-Time Adaptation by Entropy Minimization

[2:35] Neural Approximate Sufficient Statistics for Implicit Models

[2:45] Implicit Normalizing Flows

Q&As 2:55-3:08

[2:55] Q&A

(ends 3:08 PM)

3:15 p.m.

Break:

BREAK:

(ends 5:00 PM)

5 p.m.

Invited Talk:

Is My Dataset Biased?

Kate Saenko

(ends 6:00 PM)

6 p.m.

Poster Session 08 [6:00-8:00]

Posters 6:00-8:00

Anytime Sampling for Autoregressive Models via Ordered Autoencoding

Are Neural Nets Modular? Inspecting Functional Modularity Through Differentiable Weight Masks

Average-case Acceleration for Bilinear Games and Normal Matrices

Continuous Wasserstein-2 Barycenter Estimation without Minimax Optimization

Coupled Oscillatory Recurrent Neural Network (coRNN): An accurate and (gradient) stable architecture for learning long time dependencies

Differentiable Trust Region Layers for Deep Reinforcement Learning

Entropic gradient descent algorithms and wide flat minima

Faster Binary Embeddings for Preserving Euclidean Distances

Federated Learning via Posterior Averaging: A New Perspective and Practical Algorithms

IsarStep: a Benchmark for High-level Mathematical Reasoning

Iterated learning for emergent systematicity in VQA

Learning advanced mathematical computations from examples

Learning Task-General Representations with Generative Neuro-Symbolic Modeling

Learning to Represent Action Values as a Hypergraph on the Action Vertices

Long-tail learning via logit adjustment

Multiplicative Filter Networks

Multivariate Probabilistic Time Series Forecasting via Conditioned Normalizing Flows

On the Bottleneck of Graph Neural Networks and its Practical Implications

OPAL: Offline Primitive Discovery for Accelerating Offline Reinforcement Learning

Perceptual Adversarial Robustness: Defense Against Unseen Threat Models

Pre-training Text-to-Text Transformers for Concept-centric Common Sense

Self-supervised Visual Reinforcement Learning with Object-centric Representations

Watch-And-Help: A Challenge for Social Perception and Human-AI Collaboration

Witches' Brew: Industrial Scale Data Poisoning via Gradient Matching

You Only Need Adversarial Supervision for Semantic Image Synthesis

A Gradient Flow Framework For Analyzing Network Pruning

Anchor & Transform: Learning Sparse Embeddings for Large Vocabularies

Ask Your Humans: Using Human Instructions to Improve Generalization in Reinforcement Learning

Benchmarks for Deep Off-Policy Evaluation

Boost then Convolve: Gradient Boosting Meets Graph Neural Networks

Chaos of Learning Beyond Zero-sum and Coordination via Game Decompositions

Creative Sketch Generation

DARTS-: Robustly Stepping out of Performance Collapse Without Indicators

DeepAveragers: Offline Reinforcement Learning By Solving Derived Non-Parametric MDPs

Deep symbolic regression: Recovering mathematical expressions from data via risk-seeking policy gradients

Early Stopping in Deep Networks: Double Descent and How to Eliminate it

Evaluation of Neural Architectures Trained With Square Loss vs Cross-Entropy in Classification Tasks

Exploring the Uncertainty Properties of Neural Networks’ Implicit Priors in the Infinite-Width Limit

Few-Shot Bayesian Optimization with Deep Kernel Surrogates

For self-supervised learning, Rationality implies generalization, provably

Generative Time-series Modeling with Fourier Flows

Geometry-Aware Gradient Algorithms for Neural Architecture Search

Graph Information Bottleneck for Subgraph Recognition

Graph Traversal with Tensor Functionals: A Meta-Algorithm for Scalable Learning

Growing Efficient Deep Networks by Structured Continuous Sparsification

How Does Mixup Help With Robustness and Generalization?

Human-Level Performance in No-Press Diplomacy via Equilibrium Search

HyperDynamics: Meta-Learning Object and Agent Dynamics with Hypernetworks

HyperGrid Transformers: Towards A Single Model for Multiple Tasks

Interpreting Graph Neural Networks for NLP With Differentiable Edge Masking

Learning Generalizable Visual Representations via Interactive Gameplay

Learning Mesh-Based Simulation with Graph Networks

Mastering Atari with Discrete World Models

Modeling the Second Player in Distributionally Robust Optimization

More or Less: When and How to Build Convolutional Neural Network Ensembles

My Body is a Cage: the Role of Morphology in Graph-Based Incompatible Control

NAS-Bench-ASR: Reproducible Neural Architecture Search for Speech Recognition

Neural Networks for Learning Counterfactual G-Invariances from Single Environments

Neural Synthesis of Binaural Speech From Mono Audio

Optimism in Reinforcement Learning with Generalized Linear Function Approximation

Orthogonalizing Convolutional Layers with the Cayley Transform

PC2WF: 3D Wireframe Reconstruction from Raw Point Clouds

Probabilistic Numeric Convolutional Neural Networks

Property Controllable Variational Autoencoder via Invertible Mutual Dependence

Provably robust classification of adversarial examples with detection

RODE: Learning Roles to Decompose Multi-Agent Tasks

SEED: Self-supervised Distillation For Visual Representation

Sharpness-aware Minimization for Efficiently Improving Generalization

Sliced Kernelized Stein Discrepancy

Theoretical bounds on estimation error for meta-learning

Training independent subnetworks for robust prediction

TropEx: An Algorithm for Extracting Linear Terms in Deep Neural Networks

Unbiased Teacher for Semi-Supervised Object Detection

Unsupervised Audiovisual Synthesis via Exemplar Autoencoders

Variational State-Space Models for Localisation and Dense 3D Mapping in 6 DoF

(ends 8:00 PM)

8 p.m.

Oral Session 8 [8:00-11:21]

Orals 8:00-9:00

[8:00] Human-Level Performance in No-Press Diplomacy via Equilibrium Search

[8:15] Learning to Reach Goals via Iterated Supervised Learning

[8:30] Learning Invariant Representations for Reinforcement Learning without Reconstruction

[8:45] Evolving Reinforcement Learning Algorithms

Spotlights 9:00-9:10

[9:00] Image Augmentation Is All You Need: Regularizing Deep Reinforcement Learning from Pixels

Q&As 9:10-9:23

[9:10] Q&A

Orals 9:23-9:38

[9:23] Coupled Oscillatory Recurrent Neural Network (coRNN): An accurate and (gradient) stable architecture for learning long time dependencies

Spotlights 9:38-10:08

[9:38] Sequential Density Ratio Estimation for Simultaneous Optimization of Speed and Accuracy

[9:48] LambdaNetworks: Modeling long-range Interactions without Attention

[9:58] Grounded Language Learning Fast and Slow

Q&As 10:08-10:18

[10:08] Q&A

Spotlights 10:18-11:08

[10:18] Unsupervised Object Keypoint Learning using Local Spatial Predictability

[10:28] VAEBM: A Symbiosis between Variational Autoencoders and Energy-based Models

[10:38] Dynamic Tensor Rematerialization

[10:48] A Gradient Flow Framework For Analyzing Network Pruning

[10:58] Differentially Private Learning Needs Better Features (or Much More Data)

Q&As 11:08-11:21

[11:08] Q&A

(ends 11:21 PM)

11 p.m.

Expo Talk Panel:

Live Panel - Academics@ Presents: Representation Learning at Amazon

(ends 12:00 AM)

THU 6 MAY

midnight

Break:

BREAK:

(ends 1:00 AM)

1 a.m.

Outstanding Paper Session 1 [1:00-2:05]

Orals 1:00-1:45

[1:00] Neural Synthesis of Binaural Speech From Mono Audio

[1:15] EigenGame: PCA as a Nash Equilibrium

[1:30] Score-Based Generative Modeling through Stochastic Differential Equations

Spotlights 1:45-1:55

[1:45] Learning Mesh-Based Simulation with Graph Networks

Q&As 1:55-2:05

[1:55] Q&A

(ends 2:05 AM)

2 a.m.

Poster Session 09 [2:00-4:00]

Posters 2:00-4:00

Adaptive Universal Generalized PageRank Graph Neural Network

A Geometric Analysis of Deep Generative Image Models and Its Applications

AutoLRS: Automatic Learning-Rate Schedule by Bayesian Optimization on the Fly

BERTology Meets Biology: Interpreting Attention in Protein Language Models

Beyond Fully-Connected Layers with Quaternions: Parameterization of Hypercomplex Multiplications with $1/n$ Parameters

CoCo: Controllable Counterfactuals for Evaluating Dialogue State Trackers

CPR: Classifier-Projection Regularization for Continual Learning

Deep Encoder, Shallow Decoder: Reevaluating Non-autoregressive Machine Translation

Efficient Reinforcement Learning in Factored MDPs with Application to Constrained RL

Efficient Wasserstein Natural Gradients for Reinforcement Learning

Empirical or Invariant Risk Minimization? A Sample Complexity Perspective

Evaluations and Methods for Explanation through Robustness Analysis

Evolving Reinforcement Learning Algorithms

Graph-Based Continual Learning

In-N-Out: Pre-Training and Self-Training using Auxiliary Information for Out-of-Distribution Robustness

Is Attention Better Than Matrix Decomposition?

Learning and Evaluating Representations for Deep One-Class Classification

Learning to Deceive Knowledge Graph Augmented Models via Targeted Perturbation

Learning with Instance-Dependent Label Noise: A Sample Sieve Approach

Mind the Pad -- CNNs Can Develop Blind Spots

Neural Architecture Search on ImageNet in Four GPU Hours: A Theoretically Inspired Perspective

Simple Augmentation Goes a Long Way: ADRL for DNN Quantization

AdaFuse: Adaptive Temporal Fusion Network for Efficient Action Recognition

Adaptive Procedural Task Generation for Hard-Exploration Problems

AdaSpeech: Adaptive Text to Speech for Custom Voice

ALFWorld: Aligning Text and Embodied Environments for Interactive Learning

A PAC-Bayesian Approach to Generalization Bounds for Graph Neural Networks

Better Fine-Tuning by Reducing Representational Collapse

Beyond Categorical Label Representations for Image Classification

Combining Physics and Machine Learning for Network Flow Estimation

Conservative Safety Critics for Exploration

Control-Aware Representations for Model-based Reinforcement Learning

Deep Neural Tangent Kernel and Laplace Kernel Have the Same RKHS

DiffWave: A Versatile Diffusion Model for Audio Synthesis

Economic Hyperparameter Optimization With Blended Search Strategy

Effective and Efficient Vote Attack on Capsule Networks

Efficient Conformal Prediction via Cascaded Inference with Expanded Admission

Efficient Empowerment Estimation for Unsupervised Stabilization

Emergent Symbols through Binding in External Memory

Empirical Analysis of Unlabeled Entity Problem in Named Entity Recognition

Estimating informativeness of samples with Smooth Unique Information

Estimating Lipschitz constants of monotone deep equilibrium models

Evaluating the Disentanglement of Deep Generative Models through Manifold Topology

Explainable Subgraph Reasoning for Forecasting on Temporal Knowledge Graphs

Fast And Slow Learning Of Recurrent Independent Mechanisms

Filtered Inner Product Projection for Crosslingual Embedding Alignment

GANs Can Play Lottery Tickets Too

Generating Furry Cars: Disentangling Object Shape and Appearance across Multiple Domains

gradSim: Differentiable simulation for system identification and visuomotor control

Improved Autoregressive Modeling with Distribution Smoothing

Individually Fair Gradient Boosting

Influence Functions in Deep Learning Are Fragile

In Search of Lost Domain Generalization

INT: An Inequality Benchmark for Evaluating Generalization in Theorem Proving

Interactive Weak Supervision: Learning Useful Heuristics for Data Labeling

Learning Long-term Visual Dynamics with Region Proposal Interaction Networks

Learning Manifold Patch-Based Representations of Man-Made Shapes

Learning to Generate 3D Shapes with Generative Cellular Automata

Learning with Feature-Dependent Label Noise: A Progressive Approach

Local Convergence Analysis of Gradient Descent Ascent with Finite Timescale Separation

Measuring Massive Multitask Language Understanding

Meta Back-Translation

Mixed-Features Vectors and Subspace Splitting

Model-Based Visual Planning with Self-Supervised Functional Distances

Modelling Hierarchical Structure between Dialogue Policy and Natural Language Generator with Option Framework for Task-oriented Dialogue System

NBDT: Neural-Backed Decision Tree

On the Critical Role of Conventions in Adaptive Human-AI Collaboration

PolarNet: Learning to Optimize Polar Keypoints for Keypoint Based Object Detection

Protecting DNNs from Theft using an Ensemble of Diverse Models

Revisiting Dynamic Convolution via Matrix Decomposition

Robust Overfitting may be mitigated by properly learned smoothening

Task-Agnostic Morphology Evolution

Wandering within a world: Online contextualized few-shot learning

(ends 4:00 AM)

4 a.m.

Oral Session 9 [4:00-7:05]

Orals 4:00-4:15

[4:00] Improved Autoregressive Modeling with Distribution Smoothing

Spotlights 4:15-4:45

[4:15] GAN "Steerability" without optimization

[4:25] Large Scale Image Completion via Co-Modulated Generative Adversarial Networks

[4:35] Emergent Symbols through Binding in External Memory

Q&As 4:45-4:55

[4:45] Q&A

Orals 4:55-5:10

[4:55] Deformable DETR: Deformable Transformers for End-to-End Object Detection

Spotlights 5:10-6:00

[5:10] Graph-Based Continual Learning

[5:20] Understanding the role of importance weighting for deep learning

[5:30] Towards Robustness Against Natural Language Word Substitutions

[5:40] Undistillable: Making A Nasty Teacher That CANNOT teach students

[5:50] CPT: Efficient Deep Neural Network Training via Cyclic Precision

Q&As 6:00-6:15

[6:00] Q&A

Spotlights 6:15-6:55

[6:15] PlasticineLab: A Soft-Body Manipulation Benchmark with Differentiable Physics

[6:25] Regularization Matters in Policy Optimization - An Empirical Study on Continuous Control

[6:35] Regularized Inverse Reinforcement Learning

[6:45] Behavioral Cloning from Noisy Demonstrations

Q&As 6:55-7:05

[6:55] Q&A

(ends 7:05 AM)

7:15 a.m.

Break:

BREAK:

(ends 7:15 AM)

9 a.m.

Outstanding Paper Session 2 [9:00-10:05]

Orals 9:00-9:45

[9:00] Rethinking Architecture Selection in Differentiable NAS

[9:15] Complex Query Answering with Neural Link Predictors

[9:30] Optimal Rates for Averaged Stochastic Gradient Descent under Neural Tangent Kernel Regime

Spotlights 9:45-9:55

[9:45] Beyond Fully-Connected Layers with Quaternions: Parameterization of Hypercomplex Multiplications with $1/n$ Parameters

Q&As 9:55-10:05

[9:55] Q&A

(ends 10:05 AM)

10 a.m.

Poster Session 10 [10:00-12:00]

Posters 10:00-12:00

AdamP: Slowing Down the Slowdown for Momentum Optimizers on Scale-invariant Weights

Adaptive Extra-Gradient Methods for Min-Max Optimization and Games

Categorical Normalizing Flows via Continuous Transformations

Certify or Predict: Boosting Certified Robustness with Compositional Architectures

Collective Robustness Certificates: Exploiting Interdependence in Graph Neural Networks

Colorization Transformer

Continual learning in recurrent neural networks

Do not Let Privacy Overbill Utility: Gradient Embedding Perturbation for Private Learning

Genetic Soft Updates for Policy Evolution in Deep Reinforcement Learning

Grounding Language to Autonomously-Acquired Skills via Goal Generation

Incremental few-shot learning via vector quantization in deep embedded space

Influence Estimation for Generative Adversarial Networks

Interpretable Models for Granger Causality Using Self-explaining Neural Networks

Kanerva++: Extending the Kanerva Machine With Differentiable, Locally Block Allocated Latent Memory

Learning Reasoning Paths over Semantic Graphs for Video-grounded Dialogues

Network Pruning That Matters: A Case Study on Retraining Variants

Private Image Reconstruction from System Side Channels Using Generative Models

Revisiting Hierarchical Approach for Persistent Long-Term Video Prediction

Sparse Quantized Spectral Clustering

The inductive bias of ReLU networks on orthogonally separable data

Understanding the effects of data parallelism and sparsity on neural network training

Adaptive and Generative Zero-Shot Learning

A Diffusion Theory For Deep Learning Dynamics: Stochastic Gradient Descent Exponentially Favors Flat Minima

Adversarially Guided Actor-Critic

An Unsupervised Deep Learning Approach for Real-World Image Denoising

Autoregressive Entity Retrieval

Balancing Constraints and Rewards with Meta-Gradient D4PG

CausalWorld: A Robotic Manipulation Benchmark for Causal Structure and Transfer Learning

ChipNet: Budget-Aware Pruning with Heaviside Continuous Approximations

Conditional Generative Modeling via Learning the Latent Space

Contrastive Divergence Learning is a Time Reversal Adversarial Game

Contrastive Explanations for Reinforcement Learning via Embedded Self Predictions

Contrastive Learning with Adversarial Perturbations for Conditional Text Generation

Counterfactual Generative Networks

Deformable DETR: Deformable Transformers for End-to-End Object Detection

Distilling Knowledge from Reader to Retriever for Question Answering

Efficient Generalized Spherical CNNs

Efficient Inference of Flexible Interaction in Spiking-neuron Networks

Emergent Road Rules In Multi-Agent Driving Environments

Free Lunch for Few-shot Learning: Distribution Calibration

GAN "Steerability" without optimization

Go with the flow: Adaptive control for Neural ODEs

GShard: Scaling Giant Models with Conditional Computation and Automatic Sharding

Hopfield Networks is All You Need

Impact of Representation Learning in Linear Bandits

Inductive Representation Learning in Temporal Networks via Causal Anonymous Walks

IOT: Instance-wise Layer Reordering for Transformer Structures

Latent Convergent Cross Mapping

Learnable Embedding sizes for Recommender Systems

Learning continuous-time PDEs from sparse data with graph neural networks

Learning Deep Features in Instrumental Variable Regression

Learning Neural Generative Dynamics for Molecular Conformation Generation

Learning What To Do by Simulating the Past

Loss Function Discovery for Object Detection via Convergence-Simulation Driven Search

Mind the Gap when Conditioning Amortised Inference in Sequential Latent-Variable Models

Neural Topic Model via Optimal Transport

Optimal Conversion of Conventional Artificial Neural Networks to Spiking Neural Networks

Optimal Rates for Averaged Stochastic Gradient Descent under Neural Tangent Kernel Regime

Practical Massively Parallel Monte-Carlo Tree Search Applied to Molecular Design

Representation Balancing Offline Model-based Reinforcement Learning

Repurposing Pretrained Models for Robust Out-of-domain Few-Shot Learning

Retrieval-Augmented Generation for Code Summarization via Hybrid GNN

R-GAP: Recursive Gradient Attack on Privacy

Scalable Learning and MAP Inference for Nonsymmetric Determinantal Point Processes

What Matters for On-Policy Deep Actor-Critic Methods? A Large-Scale Study

(ends 12:00 PM)

noon

Oral Session 10 [12:00-2:40]

Orals 12:00-12:15

[12:00] What Matters for On-Policy Deep Actor-Critic Methods? A Large-Scale Study

Spotlights 12:15-1:05

[12:15] Winning the L2RPN Challenge: Power Grid Management via Semi-Markov Afterstate Actor-Critic

[12:25] UPDeT: Universal Multi-agent RL via Policy Decoupling with Transformers

[12:35] Quantifying Differences in Reward Functions

[12:45] Iterative Empirical Game Solving via Single Policy Best Response

[12:55] Discovering a set of policies for the worst case reward

Q&As 1:05-1:20

[1:05] Q&A

Orals 1:20-1:35

[1:20] Augmenting Physical Models with Deep Networks for Complex Dynamics Forecasting

Spotlights 1:35-2:25

[1:35] Unlearnable Examples: Making Personal Data Unexploitable

[1:45] Self-supervised Visual Reinforcement Learning with Object-centric Representations

[1:55] On Self-Supervised Image Representations for GAN Evaluation

[2:05] Retrieval-Augmented Generation for Code Summarization via Hybrid GNN

[2:15] Practical Real Time Recurrent Learning with a Sparse Approximation

Q&As 2:25-2:40

[2:25] Q&A

(ends 2:40 PM)

3 p.m.

Break:

BREAK:

(ends 5:00 PM)

5 p.m.

Invited Talk:

Soft bodied robots for human centered design of robots for everyday life

Kyu Jin Cho

(ends 6:00 PM)

6 p.m.

Poster Session 11 [6:00-8:00]

Posters 6:00-8:00

Analyzing the Expressive Power of Graph Neural Networks in a Spectral Perspective

BSQ: Exploring Bit-Level Sparsity for Mixed-Precision Neural Network Quantization

CaPC Learning: Confidential and Private Collaborative Learning

Contrastive Behavioral Similarity Embeddings for Generalization in Reinforcement Learning

Contrastive Learning with Hard Negative Samples

Deep Networks and the Multiple Manifold Problem

Dual-mode ASR: Unify and Improve Streaming ASR with Full-context Modeling

Dynamic Tensor Rematerialization

Enforcing robust control guarantees within neural network policies

Flowtron: an Autoregressive Flow-based Generative Network for Text-to-Speech Synthesis

Generative Language-Grounded Policy in Vision-and-Language Navigation with Bayes' Rule

Graph Coarsening with Neural Networks

Heating up decision boundaries: isocapacitory saturation, adversarial scenarios and generalization bounds

Hierarchical Reinforcement Learning by Discovering Intrinsic Options

Integrating Categorical Semantics into Unsupervised Domain Translation

Local Search Algorithms for Rank-Constrained Convex Optimization

Meta-learning with negative learning rates

Neural Learning of One-of-Many Solutions for Combinatorial Problems in Structured Output Spaces

On Position Embeddings in BERT

Uncertainty in Gradient Boosting via Ensembles

A Critique of Self-Expressive Deep Subspace Clustering

Adversarially-Trained Deep Nets Transfer Better: Illustration on Image Classification

Adversarial score matching and improved sampling for image generation

A Mathematical Exploration of Why Language Models Help Solve Downstream Tasks

A teacher-student framework to distill future trajectories

Bayesian Few-Shot Classification with One-vs-Each Pólya-Gamma Augmented Gaussian Processes

Blending MPC & Value Function Approximation for Efficient Reinforcement Learning

BREEDS: Benchmarks for Subpopulation Shift

C-Learning: Learning to Achieve Goals via Recursive Classification

Conditionally Adaptive Multi-Task Learning: Improving Transfer Learning in NLP Using Fewer Parameters & Less Data

Correcting experience replay for multi-agent communication

Cut out the annotator, keep the cutout: better segmentation with weak supervision

Dataset Condensation with Gradient Matching

Dataset Meta-Learning from Kernel Ridge-Regression

Deconstructing the Regularization of BatchNorm

Differentially Private Learning Needs Better Features (or Much More Data)

Directed Acyclic Graph Neural Networks

Domain-Robust Visual Imitation Learning with Mutual Information Constraints

EEC: Learning to Encode and Regenerate Images for Continual Learning

Efficient Transformers in Reinforcement Learning using Actor-Learner Distillation

EigenGame: PCA as a Nash Equilibrium

End-to-end Adversarial Text-to-Speech

End-to-End Egospheric Spatial Memory

Gradient Vaccine: Investigating and Improving Multi-task Optimization in Massively Multilingual Models

Implicit Gradient Regularization

Improving VAEs' Robustness to Adversarial Attack

Improving Zero-Shot Voice Style Transfer via Disentangled Representation Learning

Initialization and Regularization of Factorized Neural Layers

Learning Cross-Domain Correspondence for Control with Dynamics Cycle-Consistency

Learning to live with Dale's principle: ANNs with separate excitatory and inhibitory units

Learning to Recombine and Resample Data For Compositional Generalization

Learning to Set Waypoints for Audio-Visual Navigation

Lifelong Learning of Compositional Structures

Linear Last-iterate Convergence in Constrained Saddle-point Optimization

Meta-Learning of Structured Task Distributions in Humans and Machines

Model-Based Offline Planning

Multi-Class Uncertainty Calibration via Mutual Information Maximization-based Binning

Neural Spatio-Temporal Point Processes

Noise or Signal: The Role of Image Backgrounds in Object Recognition

PMI-Masking: Principled masking of correlated spans

Private Post-GAN Boosting

Rethinking Soft Labels for Knowledge Distillation: A Bias–Variance Tradeoff Perspective

Robust early-learning: Hindering the memorization of noisy labels

Understanding and Improving Encoder Layer Fusion in Sequence-to-Sequence Learning

Understanding and Improving Lexical Choice in Non-Autoregressive Translation

Variational Information Bottleneck for Effective Low-Resource Fine-Tuning

VCNet and Functional Targeted Regularization For Learning Causal Effects of Continuous Treatments

(ends 8:00 PM)

8 p.m.

Oral Session 11 [8:00-11:13]

Orals 8:00-9:00

[8:00] VCNet and Functional Targeted Regularization For Learning Causal Effects of Continuous Treatments

[8:15] SenSeI: Sensitive Set Invariance for Enforcing Individual Fairness

[8:30] When Do Curricula Work?

[8:45] Why Are Convolutional Nets More Sample-Efficient than Fully-Connected Nets?

Q&As 9:00-9:10

[9:00] Q&A

Spotlights 9:10-9:50

[9:10] Correcting experience replay for multi-agent communication

[9:20] Contrastive Behavioral Similarity Embeddings for Generalization in Reinforcement Learning

[9:30] DeepAveragers: Offline Reinforcement Learning By Solving Derived Non-Parametric MDPs

[9:40] Data-Efficient Reinforcement Learning with Self-Predictive Representations

Q&As 9:50-10:00

[9:50] Q&A

Orals 10:00-10:30

[10:00] DiffWave: A Versatile Diffusion Model for Audio Synthesis

[10:15] Self-training For Few-shot Transfer Across Extreme Task Differences

Spotlights 10:30-11:00

[10:30] A Panda? No, It's a Sloth: Slowdown Attacks on Adaptive Multi-Exit Neural Network Inference

[10:40] BUSTLE: Bottom-Up Program Synthesis Through Learning-Guided Exploration

[10:50] Disentangled Recurrent Wasserstein Autoencoder

Q&As 11:00-11:13

[11:00] Q&A

(ends 11:13 PM)

11 p.m.

Expo Talk Panel:

AI Model Efficiency Toolkit talk & demo

(ends 12:00 AM)

FRI 7 MAY

midnight

Break:

BREAK:

(ends 1:00 AM)

1 a.m.

Invited Talk:

Self-Supervision for Learning from the Bottom Up

Alexei Efros

(ends 2:00 AM)

2 a.m.

Poster Session 12 [2:00-4:00]

Posters 2:00-4:00

ANOCE: Analysis of Causal Effects with Multiple Mediators via Constrained Structural Learning

CO2: Consistent Contrast for Unsupervised Visual Representation Learning

Combining Ensembles and Data Augmentation Can Harm Your Calibration

Contrastive Syn-to-Real Generalization

Distributional Sliced-Wasserstein and Applications to Generative Modeling

Few-Shot Learning via Learning the Representation, Provably

Generative Scene Graph Networks

HeteroFL: Computation and Communication Efficient Federated Learning for Heterogeneous Clients

In Defense of Pseudo-Labeling: An Uncertainty-Aware Pseudo-label Selection Framework for Semi-Supervised Learning

Large Scale Image Completion via Co-Modulated Generative Adversarial Networks

Linear Convergent Decentralized Optimization with Compression

MARS: Markov Molecular Sampling for Multi-objective Drug Discovery

Molecule Optimization by Explainable Evolution

Neural Pruning via Growing Regularization

Non-asymptotic Confidence Intervals of Off-policy Evaluation: Primal and Dual Bounds

$i$-Mix: A Domain-Agnostic Strategy for Contrastive Representation Learning

Adapting to Reward Progressivity via Spectral Reinforcement Learning

A Design Space Study for LISTA and Beyond

A Learning Theoretic Perspective on Local Explainability

Answering Complex Open-Domain Questions with Multi-Hop Dense Retrieval

ARMOURED: Adversarially Robust MOdels using Unlabeled data by REgularizing Diversity

Async-RED: A Provably Convergent Asynchronous Block Parallel Stochastic Method using Deep Denoising Priors

Auto Seg-Loss: Searching Metric Surrogates for Semantic Segmentation

BiPointNet: Binary Neural Network for Point Clouds

Calibration of Neural Networks using Splines

Clustering-friendly Representation Learning via Instance Discrimination and Feature Decorrelation

Combining Label Propagation and Simple Models out-performs Graph Neural Networks

Convex Regularization behind Neural Reconstruction

CPT: Efficient Deep Neural Network Training via Cyclic Precision

Cross-Attentional Audio-Visual Fusion for Weakly-Supervised Action Localization

CT-Net: Channel Tensorization Network for Video Classification

DynaTune: Dynamic Tensor Program Optimization in Deep Neural Network Compilation

Estimating and Evaluating Regression Predictive Uncertainty in Deep Object Detectors

Evaluation of Similarity-based Explanations

Extreme Memorization via Scale of Initialization

Factorizing Declarative and Procedural Knowledge in Structured, Dynamical Environments

Fast and Complete: Enabling Complete Neural Network Verification with Rapid and Massively Parallel Incomplete Verifiers

Fast Geometric Projections for Local Robustness Certification

FastSpeech 2: Fast and High-Quality End-to-End Text to Speech

FedBN: Federated Learning on Non-IID Features via Local Batch Normalization

Fully Unsupervised Diversity Denoising with Convolutional Variational Autoencoders

GAN2GAN: Generative Noise Learning for Blind Denoising with Single Noisy Images

Generalization bounds via distillation

Global Convergence of Three-layer Neural Networks in the Mean Field Regime

Greedy-GQ with Variance Reduction: Finite-time Analysis and Improved Complexity

Group Equivariant Generative Adversarial Networks

Heteroskedastic and Imbalanced Deep Learning with Adaptive Regularization

How Much Over-parameterization Is Sufficient to Learn Deep ReLU Networks?

HW-NAS-Bench: Hardware-Aware Neural Architecture Search Benchmark

Isotropy in the Contextual Embedding Space: Clusters and Manifolds

Learning Energy-Based Generative Models via Coarse-to-Fine Expanding and Sampling

Learning perturbation sets for robust machine learning

Learning to Make Decisions via Submodular Regularization

Learning to Sample with Local and Global Contexts in Experience Replay Buffer

Long-tailed Recognition by Routing Diverse Distribution-Aware Experts

LowKey: Leveraging Adversarial Attacks to Protect Social Media Users from Facial Recognition

Minimum Width for Universal Approximation

Multi-Prize Lottery Ticket Hypothesis: Finding Accurate Binary Neural Networks by Pruning A Randomly Weighted Network

Multi-timescale Representation Learning in LSTM Language Models

Neural representation and generation for RNA secondary structures

Neural Thompson Sampling

No MCMC for me: Amortized sampling for fast and stable training of energy-based models

Nonseparable Symplectic Neural Networks

On the Curse of Memory in Recurrent Neural Networks: Approximation and Optimization Analysis

Prototypical Representation Learning for Relation Extraction

Randomized Automatic Differentiation

Representing Partial Programs with Blended Abstract Semantics

Self-supervised Learning from a Multi-view Perspective

Self-supervised Representation Learning with Relative Predictive Coding

Stochastic Security: Adversarial Defense Using Long-Run Dynamics of Energy-Based Models

Theoretical Analysis of Self-Training with Deep Networks on Unlabeled Data

The Recurrent Neural Tangent Kernel

When Do Curricula Work?

(ends 4:00 AM)

4 a.m.

Oral Session 12 [4:00-6:48]

Orals 4:00-4:15

[4:00] Theoretical Analysis of Self-Training with Deep Networks on Unlabeled Data

Spotlights 4:15-4:45

[4:15] Long-tailed Recognition by Routing Diverse Distribution-Aware Experts

[4:25] Self-Supervised Policy Adaptation during Deployment

[4:35] What are the Statistical Limits of Offline RL with Linear Function Approximation?

Q&As 4:45-4:55

[4:45] Q&A

Spotlights 4:55-5:45

[4:55] RMSprop converges with proper hyper-parameter

[5:05] A Good Image Generator Is What You Need for High-Resolution Video Synthesis

[5:15] Random Feature Attention

[5:25] Learning with Feature-Dependent Label Noise: A Progressive Approach

[5:35] Sparse Quantized Spectral Clustering

Q&As 5:45-5:58

[5:45] Q&A

Spotlights 5:58-6:38

[5:58] Learning a Latent Simplex in Input Sparsity Time

[6:08] Topology-Aware Segmentation Using Discrete Morse Theory

[6:18] MARS: Markov Molecular Sampling for Multi-objective Drug Discovery

[6:28] Distributional Sliced-Wasserstein and Applications to Generative Modeling

Q&As 6:38-6:48

[6:38] Q&A

(ends 6:48 AM)

7 a.m.

Break:

BREAK:

(ends 7:00 AM)

Remarks:

Closing Remarks

(ends 8:00 AM)

11:30 a.m.

Workshop:

Science and Engineering of Deep Learning

(ends 8:30 PM)

12:30 p.m.

Workshop:

Neural Compression: From Information Theory to Applications

(ends 10:30 PM)

1:45 p.m.

Workshop:

Hardware-Aware Efficient Training of Deep Learning Models

(ends 9:00 PM)

2 p.m.

Workshop:

Geometric and Topological Representation Learning

(ends 1:00 AM)

Workshop:

S2D-OLAD: From shallow to deep, overcoming limited and adverse data

(ends 12:00 AM)

2:15 p.m.

Workshop:

Beyond Static Papers: Rethinking How We Share Scientific Understanding in ML

(ends 7:00 PM)

2:45 p.m.

Workshop:

Self-Supervision for Reinforcement Learning

(ends 1:30 AM)

2:50 p.m.

Workshop:

Energy-Based Models: Current Perspectives, Challenges, and Opportunities

(ends 1:10 AM)

2:55 p.m.

Workshop:

AI for Public Health

(ends 12:00 AM)

Workshop:

The Role of Mathematical Reasoning in General Artificial Intelligence

(ends 11:35 PM)

3 p.m.

Workshop:

AIMOCC -- AI: Modeling Oceans and Climate Change

(ends 7:00 PM)

Workshop:

A Roadmap to Never-Ending RL

(ends 10:55 PM)

Workshop:

Workshop on Neural Architecture Search

(ends 12:00 AM)

Break:

BREAK:

(ends 3:00 PM)

3:30 p.m.

Workshop:

How Can Findings About The Brain Improve AI Systems?

(ends 2:30 AM)

3:45 p.m.

Workshop:

Responsible AI (RAI)

(ends 4:00 AM)

4 p.m.

Workshop:

2nd Workshop on Practical ML for Developing Countries: Learning Under Limited/low Resource Scenarios

(ends 9:15 PM)

Workshop:

Synthetic Data Generation: Quality, Privacy, Bias

(ends 11:30 PM)

Workshop:

Neural Conversational AI: Bridging the Gap Between Research and Real World (NeuCAIR)

(ends 5:30 AM)

Workshop:

Generalization beyond the training distribution in brains and machines

(ends 12:55 AM)

Workshop:

Workshop on Weakly Supervised Learning

(ends 3:00 AM)

Workshop:

Workshop on Learning to Learn

(ends 9:45 PM)

4:45 p.m.

Workshop:

Workshop on Enormous Language Models: Perspectives and Benchmarks

(ends 2:00 AM)

4:55 p.m.

Workshop:

ICLR 2021 Workshop on Embodied Multimodal Learning (EML)

(ends 12:35 AM)

5 p.m.

Workshop:

Robust and reliable machine learning in the real world

(ends 2:05 AM)

5:30 p.m.

Workshop:

Workshop on Distributed and Private Machine Learning

(ends 10:00 PM)

5:45 p.m.

Workshop:

Security and Safety in Machine Learning Systems

(ends 3:00 AM)

Workshop:

Machine Learning for Preventing and Combating Pandemics

(ends 2:50 AM)

Workshop:

Deep Learning for Simulation

(ends 2:00 AM)

11 p.m.

Break:

BREAK:

(ends 11:00 PM)