ICLR 2022 Schedule

Filter Events

MON 25 APR

7 a.m.

Remarks:

Opening Remarks

(ends 7:55 AM)

8 a.m.

Invited Talk:

Leveraging AI for Science

Pushmeet Kohli

(ends 9:15 AM)

9:30 a.m.

Poster Session 1 [9:30-11:30]

PF-GNN: Differentiable particle filtering based approximation of universal graph representations

Who Is Your Right Mixup Partner in Positive and Unlabeled Learning

Constrained Policy Optimization via Bayesian World Models

The Role of Permutation Invariance in Linear Mode Connectivity of Neural Networks

Understanding and Improving Graph Injection Attack by Promoting Unnoticeability

Evidential Turing Processes

Generalisation in Lifelong Reinforcement Learning through Logical Composition

Expressivity of Emergent Languages is a Trade-off between Contextual Complexity and Unpredictability

Learning Fast, Learning Slow: A General Continual Learning Method based on Complementary Learning System

Learning Temporally Causal Latent Processes from General Temporal Data

Bootstrapping Semantic Segmentation with Regional Contrast

Dropout Q-Functions for Doubly Efficient Reinforcement Learning

Training Structured Neural Networks Through Manifold Identification and Variance Reduction

Scale Mixtures of Neural Network Gaussian Processes

PriorGrad: Improving Conditional Denoising Diffusion Models with Data-Dependent Adaptive Prior

MetaShift: A Dataset of Datasets for Evaluating Contextual Distribution Shifts and Training Conflicts

Encoding Weights of Irregular Sparsity for Fixed-to-Fixed Model Compression

Source-Free Adaptation to Measurement Shift via Bottom-Up Feature Restoration

Hybrid Local SGD for Federated Learning with Heterogeneous Communications

Sample and Computation Redistribution for Efficient Face Detection

Semi-relaxed Gromov-Wasserstein divergence and applications on graphs

Bayesian Neural Network Priors Revisited

Perceiver IO: A General Architecture for Structured Inputs & Outputs

Topological Graph Neural Networks

Energy-Based Learning for Cooperative Games, with Applications to Valuation Problems in Machine Learning

Progressive Distillation for Fast Sampling of Diffusion Models

Signing the Supermask: Keep, Hide, Invert

Generative Modeling with Optimal Transport Maps

Autoregressive Diffusion Models

Imbedding Deep Neural Networks

Top-label calibration and multiclass-to-binary reductions

High Probability Generalization Bounds with Fast Rates for Minimax Problems

Steerable Partial Differential Operators for Equivariant Neural Networks

Prospect Pruning: Finding Trainable Weights at Initialization using Meta-Gradients

Measuring the Interpretability of Unsupervised Representations via Quantized Reversed Probing

The Three Stages of Learning Dynamics in High-dimensional Kernel Methods

The Role of Pretrained Representations for the OOD Generalization of RL Agents

Deconstructing the Inductive Biases of Hamiltonian Neural Networks

End-to-End Learning of Probabilistic Hierarchies on Graphs

Spanning Tree-based Graph Generation for Molecules

8-bit Optimizers via Block-wise Quantization

IntSGD: Adaptive Floatless Compression of Stochastic Gradients

On the Existence of Universal Lottery Tickets

On Incorporating Inductive Biases into VAEs

Controlling the Complexity and Lipschitz Constant improves Polynomial Nets

Stability Regularization for Discrete Representation Learning

Collapse by Conditioning: Training Class-conditional GANs with Limited Data

Boosted Curriculum Reinforcement Learning

The Unreasonable Effectiveness of Random Pruning: Return of the Most Naive Baseline for Sparse Training

Boosting the Certified Robustness of L-infinity Distance Nets

Efficient Neural Causal Discovery without Acyclicity Constraints

Unsupervised Disentanglement with Tensor Product Representations on the Torus

Learning to Generalize across Domains on Single Test Samples

Hierarchical Variational Memory for Few-shot Learning Across Domains

Learning Long-Term Reward Redistribution via Randomized Return Decomposition

Analytic-DPM: an Analytic Estimate of the Optimal Reverse Variance in Diffusion Probabilistic Models

Goal-Directed Planning via Hindsight Experience Replay

Learning Synthetic Environments and Reward Networks for Reinforcement Learning

Graph-based Nearest Neighbor Search in Hyperbolic Spaces

switch-GLAT: Multilingual Parallel Machine Translation Via Code-Switch Decoder

Variational Neural Cellular Automata

Random matrices in service of ML footprint: ternary random features with no performance loss

Adversarial Retriever-Ranker for Dense Text Retrieval

Joint Shapley values: a measure of joint feature importance

Non-Parallel Text Style Transfer with Self-Parallel Supervision

Trans-Encoder: Unsupervised sentence-pair modelling through self- and mutual-distillations

SUMNAS: Supernet with Unbiased Meta-Features for Neural Architecture Search

A Fine-Grained Analysis on Distribution Shift

Deep Ensembling with No Overhead for either Training or Testing: The All-Round Blessings of Dynamic Sparsity

SQuant: On-the-Fly Data-Free Quantization via Diagonal Hessian Approximation

Memory Replay with Data Compression for Continual Learning

Anytime Dense Prediction with Confidence Adaptivity

Using Graph Representation Learning with Schema Encoders to Measure the Severity of Depressive Symptoms

(ends 11:30 AM)

11 a.m.

4 p.m.

Invited Talk:

Beyond interpretability: developing a language to shape our relationships with AI

Been Kim

(ends 5:15 PM)

5:30 p.m.

Poster Session 2 [5:30-7:30]

How unlabeled data improve generalization in self-training? A one-hidden-layer theoretical analysis

Learning Transferable Reward for Query Object Localization with Policy Adaptation

The Spectral Bias of Polynomial Neural Networks

A Generalized Weighted Optimization Method for Computational Learning and Inversion

Open-vocabulary Object Detection via Vision and Language Knowledge Distillation

Eliminating Sharp Minima from SGD with Truncated Heavy-tailed Noise

Critical Points in Quantum Generative Models

Gradient Importance Learning for Incomplete Observations

Learning to Guide and to be Guided in the Architect-Builder Problem

Understanding Intrinsic Robustness Using Label Uncertainty

Learn Locally, Correct Globally: A Distributed Algorithm for Training Graph Neural Networks

Efficient Self-supervised Vision Transformers for Representation Learning

DISSECT: Disentangled Simultaneous Explanations via Concept Traversals

PAC Prediction Sets Under Covariate Shift

Fast Model Editing at Scale

Task Affinity with Maximum Bipartite Matching in Few-Shot Learning

Hybrid Memoised Wake-Sleep: Approximate Inference at the Discrete-Continuous Interface

$\mathrm{SO}(2)$-Equivariant Reinforcement Learning

On Predicting Generalization using GANs

Which Shortcut Cues Will DNNs Choose? A Study from the Parameter-Space Perspective

Tackling the Generative Learning Trilemma with Denoising Diffusion GANs

Towards Better Understanding and Better Generalization of Low-shot Classification in Histology Images with Contrastive Learning

Illiterate DALL-E Learns to Compose

Graph-Augmented Normalizing Flows for Anomaly Detection of Multiple Time Series

Salient ImageNet: How to discover spurious features in Deep Learning?

Distributional Reinforcement Learning with Monotonic Splines

MaGNET: Uniform Sampling from Deep Generative Network Manifolds Without Retraining

Environment Predictive Coding for Visual Navigation

Revisiting Design Choices in Offline Model Based Reinforcement Learning

A Tale of Two Flows: Cooperative Learning of Langevin Flow and Normalizing Flow Toward Energy-Based Model

Vision-Based Manipulators Need to Also See from Their Hands

Chunked Autoregressive GAN for Conditional Waveform Synthesis

An Experimental Design Perspective on Model-Based Reinforcement Learning

Quadtree Attention for Vision Transformers

Optimizer Amalgamation

Information-theoretic Online Memory Selection for Continual Learning

How Much Can CLIP Benefit Vision-and-Language Tasks?

Self-Supervised Graph Neural Networks for Improved Electroencephalographic Seizure Analysis

Scalable Sampling for Nonsymmetric Determinantal Point Processes

Coordination Among Neural Modules Through a Shared Global Workspace

MCMC Should Mix: Learning Energy-Based Model with Neural Transport Latent Space MCMC

Fast AdvProp

Towards Empirical Sandwich Bounds on the Rate-Distortion Function

Learning Efficient Image Super-Resolution Networks via Structure-Regularized Pruning

Real-Time Neural Voice Camouflage

When Vision Transformers Outperform ResNets without Pre-training or Strong Data Augmentations

Hierarchical Few-Shot Imitation with Skill Transition Models

Spherical Message Passing for 3D Molecular Graphs

Anisotropic Random Feature Regression in High Dimensions

Language-driven Semantic Segmentation

Fast topological clustering with Wasserstein distance

Stochastic Training is Not Necessary for Generalization

Modular Lifelong Reinforcement Learning via Neural Composition

Exploring the Limits of Large Scale Pre-training

Mind the Gap: Domain Gap Control for Single Shot Domain Adaptation for Generative Adversarial Networks

From Stars to Subgraphs: Uplifting Any GNN with Local Structure Awareness

Escaping limit cycles: Global convergence for constrained nonconvex-nonconcave minimax problems

Multiset-Equivariant Set Prediction with Approximate Implicit Differentiation

Offline Reinforcement Learning with Implicit Q-Learning

Continual Learning with Filter Atom Swapping

MobileViT: Light-weight, General-purpose, and Mobile-friendly Vision Transformer

FairCal: Fairness Calibration for Face Verification

Should I Run Offline Reinforcement Learning or Behavioral Cloning?

Group-based Interleaved Pipeline Parallelism for Large-scale DNN Training

EXACT: Scalable Graph Neural Networks Training via Extreme Activation Compression

Peek-a-Boo: What (More) is Disguised in a Randomly Weighted Neural Network, and How to Find It Efficiently

Anti-Oversmoothing in Deep Vision Transformers via the Fourier Domain Analysis: From Theory to Practice

NAS-Bench-Suite: NAS Evaluation is (Now) Surprisingly Easy

AEVA: Black-box Backdoor Detection Using Adversarial Extreme Value Analysis

Privacy Implications of Shuffling

COPA: Certifying Robust Policies for Offline Reinforcement Learning against Poisoning Attacks

Graph Condensation for Graph Neural Networks

Auto-Transfer: Learning to Route Transferable Representations

Adversarial Support Alignment

Incremental False Negative Detection for Contrastive Learning

Hyperparameter Tuning with Renyi Differential Privacy

Effect of scale on catastrophic forgetting in neural networks

EigenGame Unloaded: When playing games is better than optimizing

The Inductive Bias of In-Context Learning: Rethinking Pretraining Example Design

Sound and Complete Neural Network Repair with Minimality and Locality Guarantees

Bridging the Gap: Providing Post-Hoc Symbolic Explanations for Sequential Decision-Making Problems with Inscrutable Representations

Score-Based Generative Modeling with Critically-Damped Langevin Diffusion

Noisy Feature Mixup

Hindsight is 20/20: Leveraging Past Traversals to Aid 3D Perception

DEGREE: Decomposition Based Explanation for Graph Neural Networks

NodePiece: Compositional and Parameter-Efficient Representations of Large Knowledge Graphs

Efficiently Modeling Long Sequences with Structured State Spaces

What Makes Better Augmentation Strategies? Augment Difficult but Not too Different

Learning by Directional Gradient Descent

Relating transformers to models and neural representations of the hippocampal formation

Near-optimal Offline Reinforcement Learning with Linear Representation: Leveraging Variance Information with Pessimism

Audio Lottery: Speech Recognition Made Ultra-Lightweight, Noise-Robust, and Transferable

It Takes Four to Tango: Multiagent Self Play for Automatic Curriculum Generation

Network Augmentation for Tiny Deep Learning

DictFormer: Tiny Transformer with Shared Dictionary

Transform2Act: Learning a Transform-and-Control Policy for Efficient Agent Design

Toward Faithful Case-based Reasoning through Learning Prototypes in a Nearest Neighbor-friendly Space.

Deep AutoAugment

Chemical-Reaction-Aware Molecule Representation Learning

Learning Audio-Visual Speech Representation by Masked Multimodal Cluster Prediction

Igeood: An Information Geometry Approach to Out-of-Distribution Detection

Visual Correspondence Hallucination

Continuous-Time Meta-Learning with Forward Mode Differentiation

How to deal with missing data in supervised deep learning?

Scarf: Self-Supervised Contrastive Learning using Random Feature Corruption

Fair Normalizing Flows

THOMAS: Trajectory Heatmap Output with learned Multi-Agent Sampling

Contextualized Scene Imagination for Generative Commonsense Reasoning

W-CTC: a Connectionist Temporal Classification Loss with Wild Cards

CROP: Certifying Robust Policies for Reinforcement Learning through Functional Smoothing

Learning Hierarchical Structures with Differentiable Nondeterministic Stacks

Controlling Directions Orthogonal to a Classifier

Neural Relational Inference with Node-Specific Information

Sqrt(d) Dimension Dependence of Langevin Monte Carlo

Mapping Language Models to Grounded Conceptual Spaces

RegionViT: Regional-to-Local Attention for Vision Transformers

Transition to Linearity of Wide Neural Networks is an Emerging Property of Assembling Weak Models

Givens Coordinate Descent Methods for Rotation Matrix Learning in Trainable Embedding Indexes

Actor-critic is implicitly biased towards high entropy optimal policies

Spatial Graph Attention and Curiosity-driven Policy for Antiviral Drug Discovery

(ends 7:30 PM)

7 p.m.

Affinity Social:

LatinX in AI (LXAI) Social @ ICLR 2022

(ends 11:59 PM)

Affinity Social:

WiML Virtual Social

(ends 9:00 PM)

TUE 26 APR

midnight

Oral 1: AI Applications [12:00-1:30]

Oral s 12:00-1:30

[12:00] Language modeling via stochastic processes

[12:15] MIDI-DDSP: Detailed Control of Musical Performance via Hierarchical Modeling

[12:30] Real-Time Neural Voice Camouflage

[12:45] ProtoRes: Proto-Residual Network for Pose Authoring via Learned Inverse Kinematics

[1:00] Open-Set Recognition: A Good Closed-Set Classifier is All You Need

[1:15] Vision-Based Manipulators Need to Also See from Their Hands

(ends 1:30 AM)

Oral 1: Learning in the wild, Reinforcement learning [12:00-1:30]

Oral s 12:00-1:30

[12:00] Hyperparameter Tuning with Renyi Differential Privacy

[12:15] PiCO: Contrastive Label Disambiguation for Partial Label Learning

[12:30] Poisoning and Backdooring Contrastive Learning

[12:45] Transform2Act: Learning a Transform-and-Control Policy for Efficient Agent Design

[1:00] The Information Geometry of Unsupervised Reinforcement Learning

[1:15] Provably Filtering Exogenous Distractors using Multistep Inverse Dynamics

(ends 1:30 AM)

1:30 a.m.

Poster Session 3 [1:30-3:30]

Neural Processes with Stochastic Attention: Paying more attention to the context dataset

Geometric Transformers for Protein Interface Contact Prediction

Iterated Reasoning with Mutual Information in Cooperative and Byzantine Decentralized Teaming

Effective Model Sparsification by Scheduled Grow-and-Prune Methods

MoReL: Multi-omics Relational Learning

Generalized Decision Transformer for Offline Hindsight Information Matching

Equivariant Transformers for Neural Network based Molecular Potentials

Path Integral Sampler: A Stochastic Control Approach For Sampling

GNN is a Counter? Revisiting GNN for Question Answering

BiBERT: Accurate Fully Binarized BERT

Towards Deepening Graph Neural Networks: A GNTK-based Optimization Perspective

Task Relatedness-Based Generalization Bounds for Meta Learning

Demystifying Batch Normalization in ReLU Networks: Equivalent Convex Optimization Models and Implicit Regularization

Reinforcement Learning with Sparse Rewards using Guidance from Offline Demonstration

Generative Models as a Data Source for Multiview Representation Learning

Focus on the Common Good: Group Distributional Robustness Follows

Chaos is a Ladder: A New Theoretical Understanding of Contrastive Learning via Augmentation Overlap

Learning Generalizable Representations for Reinforcement Learning via Adaptive Meta-learner of Behavioral Similarities

Temporal Alignment Prediction for Supervised Representation Learning and Few-Shot Sequence Classification

Emergent Communication at Scale

ConFeSS: A Framework for Single Source Cross-Domain Few-Shot Learning

Towards Continual Knowledge Learning of Language Models

Continual Normalization: Rethinking Batch Normalization for Online Continual Learning

GeoDiff: A Geometric Diffusion Model for Molecular Conformation Generation

FastSHAP: Real-Time Shapley Value Estimation

An Unconstrained Layer-Peeled Perspective on Neural Collapse

Efficient and Differentiable Conformal Prediction with General Function Classes

An Information Fusion Approach to Learning with Instance-Dependent Label Noise

Differentiable Gradient Sampling for Learning Implicit 3D Scene Reconstructions from a Single Image

Zero-CL: Instance and Feature decorrelation for negative-free symmetric contrastive learning

Safe Neurosymbolic Learning with Differentiable Symbolic Execution

Ancestral protein sequence reconstruction using a tree-structured Ornstein-Uhlenbeck variational autoencoder

Rethinking Adversarial Transferability from a Data Distribution Perspective

Retriever: Learning Content-Style Representation as a Token-Level Bipartite Graph

GraphENS: Neighbor-Aware Ego Network Synthesis for Class-Imbalanced Node Classification

Fast Generic Interaction Detection for Model Interpretability and Compression

Compositional Training for End-to-End Deep AUC Maximization

Representational Continuity for Unsupervised Continual Learning

What Do We Mean by Generalization in Federated Learning?

Self-Supervision Enhanced Feature Selection with Correlated Gates

Iterative Refinement Graph Neural Network for Antibody Sequence-Structure Co-design

Surreal-GAN:Semi-Supervised Representation Learning via GAN for uncovering heterogeneous disease-related imaging patterns

GreaseLM: Graph REASoning Enhanced Language Models

Objects in Semantic Topology

Learning Disentangled Representation by Exploiting Pretrained Generative Models: A Contrastive Learning View

A Relational Intervention Approach for Unsupervised Dynamics Generalization in Model-Based Reinforcement Learning

ViDT: An Efficient and Effective Fully Transformer-based Object Detector

Do Not Escape From the Manifold: Discovering the Local Coordinates on the Latent Space of GANs

Understanding the Role of Self Attention for Efficient Speech Recognition

Towards Evaluating the Robustness of Neural Networks Learned by Transduction

Online Coreset Selection for Rehearsal-based Continual Learning

On Evaluation Metrics for Graph Generative Models

ViTGAN: Training GANs with Vision Transformers

Meta Discovery: Learning to Discover Novel Classes given Very Limited Data

Neural Models for Output-Space Invariance in Combinatorial Problems

Variational oracle guiding for reinforcement learning

VAT-Mart: Learning Visual Action Trajectory Proposals for Manipulating 3D ARTiculated Objects

Automated Self-Supervised Learning for Graphs

UniFormer: Unified Transformer for Efficient Spatial-Temporal Representation Learning

Target-Side Input Augmentation for Sequence to Sequence Generation

Language modeling via stochastic processes

Neural Program Synthesis with Query

Hot-Refresh Model Upgrades with Regression-Free Compatible Training in Image Retrieval

On the Learning and Learnability of Quasimetrics

A Unified Contrastive Energy-based Model for Understanding the Generative Ability of Adversarial Training

Weighted Training for Cross-Task Learning

Probabilistic Implicit Scene Completion

AdaRL: What, Where, and How to Adapt in Transfer Reinforcement Learning

Trigger Hunting with a Topological Prior for Trojan Detection

Tuformer: Data-driven Design of Transformers for Improved Generalization or Efficiency

Generating Videos with Dynamics-aware Implicit Generative Adversarial Networks

Continual Learning with Recursive Gradient Optimization

The Close Relationship Between Contrastive Learning and Meta-Learning

FedBABU: Toward Enhanced Representation for Federated Image Classification

Conditional Image Generation by Conditioning Variational Auto-Encoders

Training invariances and the low-rank phenomenon: beyond linear networks

On Robust Prefix-Tuning for Text Classification

Regularized Autoencoders for Isometric Representation Learning

Optimal ANN-SNN Conversion for High-accuracy and Ultra-low-latency Spiking Neural Networks

Neural Contextual Bandits with Deep Representation and Shallow Exploration

Learning curves for continual learning in neural networks: Self-knowledge transfer and forgetting

Decoupled Adaptation for Cross-Domain Object Detection

Resonance in Weight Space: Covariate Shift Can Drive Divergence of SGD with Momentum

Taming Sparsely Activated Transformer with Stochastic Experts

On Non-Random Missing Labels in Semi-Supervised Learning

Learning Vision-Guided Quadrupedal Locomotion End-to-End with Cross-Modal Transformers

Bayesian Modeling and Uncertainty Quantification for Learning to Optimize: What, Why, and How

Data-Efficient Graph Grammar Learning for Molecular Generation

Pixelated Butterfly: Simple and Efficient Sparse training for Neural Network Models

Is High Variance Unavoidable in RL? A Case Study in Continuous Control

Unsupervised Discovery of Object Radiance Fields

(ends 3:30 AM)

3 a.m.

Affinity Social:

Queer in AI Social @ ICLR 2022

(duration 1.0 hr)

8 a.m.

Oral 2: Structured learning [8:00-9:30]

Oral s 8:00-9:30

[8:00] Understanding over-squashing and bottlenecks on graphs via curvature

[8:15] Efficiently Modeling Long Sequences with Structured State Spaces

[8:30] Neural Structured Prediction for Inductive Node Classification

[8:45] A New Perspective on "How Graph Neural Networks Go Beyond Weisfeiler-Lehman?"

[9:00] CycleMLP: A MLP-like Architecture for Dense Prediction

[9:15] Variational Inference for Discriminative Learning with Generative Modeling of Feature Incompletion

(ends 9:30 AM)

Oral 2: Understanding Deep Learning [8:00-9:45]

Oral s 8:00-9:45

[8:00] Expressiveness and Approximation Properties of Graph Neural Networks

[8:15] Neural Collapse Under MSE Loss: Proximity to and Dynamics on the Central Path

[8:30] Learning Strides in Convolutional Neural Networks

[8:45] The Hidden Convex Optimization Landscape of Regularized Two-Layer ReLU Networks: an Exact Characterization of Optimal Solutions

[9:00] Minibatch vs Local SGD with Shuffling: Tight Convergence Bounds and Beyond

[9:15] DISCOVERING AND EXPLAINING THE REPRESENTATION BOTTLENECK OF DNNS

[9:30] Representational Continuity for Unsupervised Continual Learning

(ends 9:45 AM)

Oral 2: AI applications [8:00-9:30]

Oral s 8:00-9:30

[8:00] Filtered-CoPhy: Unsupervised Learning of Counterfactual Physics in Pixel Space

[8:15] Non-Transferable Learning: A New Approach for Model Ownership Verification and Applicability Authorization

[8:30] Data-Efficient Graph Grammar Learning for Molecular Generation

[8:45] iLQR-VAE : control-based learning of input-driven dynamics with applications to neural data

[9:00] Einops: Clear and Reliable Tensor Manipulations with Einstein-like Notation

[9:15] StyleAlign: Analysis and Applications of Aligned StyleGAN Models

(ends 9:30 AM)

9:30 a.m.

Poster Session 4 [9:30-11:30]

Divergence-aware Federated Self-Supervised Learning

Trust Region Policy Optimisation in Multi-Agent Reinforcement Learning

Self-ensemble Adversarial Training for Improved Robustness

L0-Sparse Canonical Correlation Analysis

Learning meta-features for AutoML

Learning Optimal Conformal Classifiers

PiCO: Contrastive Label Disambiguation for Partial Label Learning

Measuring CLEVRness: Black-box Testing of Visual Reasoning Models

Large-Scale Representation Learning on Graphs via Bootstrapping

DeSKO: Stability-Assured Robust Control with a Deep Stochastic Koopman Operator

Neural Variational Dropout Processes

On the Connection between Local Attention and Dynamic Depth-wise Convolution

Sound Adversarial Audio-Visual Navigation

Zero Pixel Directional Boundary by Vector Transform

NeuPL: Neural Population Learning

A generalization of the randomized singular value decomposition

Explainable GNN-Based Models over Knowledge Graphs

Handling Distribution Shifts on Graphs: An Invariance Perspective

Self-Supervised Inference in State-Space Models

StyleAlign: Analysis and Applications of Aligned StyleGAN Models

Transformers Can Do Bayesian Inference

Filtered-CoPhy: Unsupervised Learning of Counterfactual Physics in Pixel Space

Dealing with Non-Stationarity in MARL via Trust-Region Decomposition

GPT-Critic: Offline Reinforcement Learning for End-to-End Task-Oriented Dialogue Systems

Autonomous Learning of Object-Centric Abstractions for High-Level Planning

Differentiable Prompt Makes Pre-trained Language Models Better Few-shot Learners

Do We Need Anisotropic Graph Neural Networks?

Fooling Explanations in Text Classifiers

Learning Strides in Convolutional Neural Networks

Distributionally Robust Models with Parametric Likelihood Ratios

Unrolling PALM for Sparse Semi-Blind Source Separation

A Neural Tangent Kernel Perspective of Infinite Tree Ensembles

Offline Neural Contextual Bandits: Pessimism, Optimization and Generalization

A Conditional Point Diffusion-Refinement Paradigm for 3D Point Cloud Completion

Generative Principal Component Analysis

DAB-DETR: Dynamic Anchor Boxes are Better Queries for DETR

Enabling Arbitrary Translation Objectives with Adaptive Tree Search

Resolving Training Biases via Influence-based Data Relabeling

Visual Representation Learning over Latent Domains

Revisiting Over-smoothing in BERT from the Perspective of Graph

Label-Efficient Semantic Segmentation with Diffusion Models

iLQR-VAE : control-based learning of input-driven dynamics with applications to neural data

DemoDICE: Offline Imitation Learning with Supplementary Imperfect Demonstrations

VAE Approximation Error: ELBO and Exponential Families

COptiDICE: Offline Constrained Reinforcement Learning via Stationary Distribution Correction Estimation

Augmented Sliced Wasserstein Distances

Rethinking Supervised Pre-Training for Better Downstream Transferring

Learning Efficient Online 3D Bin Packing on Packing Configuration Trees

Delaunay Component Analysis for Evaluation of Data Representations

Evolutionary Diversity Optimization with Clustering-based Selection for Reinforcement Learning

Relational Learning with Variational Bayes

CoST: Contrastive Learning of Disentangled Seasonal-Trend Representations for Time Series Forecasting

Simple GNN Regularisation for 3D Molecular Property Prediction and Beyond

On the approximation properties of recurrent encoder-decoder architectures

miniF2F: a cross-system benchmark for formal Olympiad-level mathematics

Amortized Implicit Differentiation for Stochastic Bilevel Optimization

BDDM: Bilateral Denoising Diffusion Models for Fast and High-Quality Speech Synthesis

Inverse Online Learning: Understanding Non-Stationary and Reactionary Policies

Optimal Transport for Long-Tailed Recognition with Learnable Cost Matrix

OBJECT DYNAMICS DISTILLATION FOR SCENE DECOMPOSITION AND REPRESENTATION

Leveraging Automated Unit Tests for Unsupervised Code Translation

Top-N: Equivariant Set and Graph Generation without Exchangeability

A First-Occupancy Representation for Reinforcement Learning

Learning Graphon Mean Field Games and Approximate Nash Equilibria

Contrastive Fine-grained Class Clustering via Generative Adversarial Networks

A Program to Build E(N)-Equivariant Steerable CNNs

MAML is a Noisy Contrastive Learner in Classification

Visual Representation Learning Does Not Generalize Strongly Within the Same Domain

In a Nutshell, the Human Asked for This: Latent Goals for Following Temporal Specifications

On Lottery Tickets and Minimal Task Representations in Deep Reinforcement Learning

A New Perspective on "How Graph Neural Networks Go Beyond Weisfeiler-Lehman?"

Phase Collapse in Neural Networks

Learning the Dynamics of Physical Systems from Sparse Observations with Finite Element Networks

Permutation Compressors for Provably Faster Distributed Nonconvex Optimization

Constrained Physical-Statistics Models for Dynamical System Identification and Prediction

Global Convergence of Multi-Agent Policy Gradient in Markov Potential Games

Scattering Networks on the Sphere for Scalable and Rotationally Equivariant Spherical CNNs

Imitation Learning by Reinforcement Learning

Learning to Downsample for Segmentation of Ultra-High Resolution Images

The Evolution of Uncertainty of Learning in Games

Unsupervised Vision-Language Grammar Induction with Shared Structure Modeling

Gradient Step Denoiser for convergent Plug-and-Play

(ends 11:30 AM)

4 p.m.

Invited Talk:

Representation Learning in the Global South: Societal Considerations-Fairness, Safety and Privacy

John Amuasi

(ends 5:15 PM)

5:30 p.m.

Poster Session 5 [5:30-7:30]

ZeroFL: Efficient On-Device Training for Federated Learning with Local Sparsity

Know Your Action Set: Learning Action Relations for Reinforcement Learning

You are AllSet: A Multiset Function Framework for Hypergraph Neural Networks

Label Encoding for Regression Networks

Improving Mutual Information Estimation with Annealed and Energy-Based Bounds

Optimizing Neural Networks with Gradient Lexicase Selection

Post-Training Detection of Backdoor Attacks for Two-Class and Multi-Attack Scenarios

Efficient Computation of Deep Nonlinear Infinite-Width Neural Networks that Learn Features

Wisdom of Committees: An Overlooked Approach To Faster and More Accurate Models

Interpretable Unsupervised Diversity Denoising and Artefact Removal

Tighter Sparse Approximation Bounds for ReLU Neural Networks

Policy Gradients Incorporating the Future

Byzantine-Robust Learning on Heterogeneous Datasets via Bucketing

RISP: Rendering-Invariant State Predictor with Differentiable Simulation and Rendering for Cross-Domain Parameter Estimation

VC dimension of partially quantized neural networks in the overparametrized regime

Pretraining Text Encoders with Adversarial Mixture of Training Signal Generators

Graph Neural Network Guided Local Search for the Traveling Salesperson Problem

Generalized Kernel Thinning

Policy Smoothing for Provably Robust Reinforcement Learning

Transfer RL across Observation Feature Spaces via Model-Based Regularization

PI3NN: Out-of-distribution-aware Prediction Intervals from Three Neural Networks

Variational methods for simulation-based inference

Promoting Saliency From Depth: Deep Unsupervised RGB-D Saliency Detection

Solving Inverse Problems in Medical Imaging with Score-Based Generative Models

Modeling Label Space Interactions in Multi-label Classification using Box Embeddings

No Parameters Left Behind: Sensitivity Guided Adaptive Learning Rate for Training Large Transformer Models

Frequency-aware SGD for Efficient Embedding Learning with Provable Benefits

Towards Model Agnostic Federated Learning Using Knowledge Distillation

Closed-form Sample Probing for Learning Generative Models in Zero-shot Learning

Minimax Optimality (Probably) Doesn't Imply Distribution Learning for GANs

Learning Value Functions from Undirected State-only Experience

Data Efficient Language-Supervised Zero-Shot Recognition with Optimal Transport Distillation

Provably convergent quasistatic dynamics for mean-field two-player zero-sum games

Temporal Efficient Training of Spiking Neural Network via Gradient Re-weighting

Fairness Guarantees under Demographic Shift

Multitask Prompted Training Enables Zero-Shot Task Generalization

Learning Neural Contextual Bandits through Perturbed Rewards

Doubly Adaptive Scaled Algorithm for Machine Learning Using Second-Order Information

Rethinking Network Design and Local Geometry in Point Cloud: A Simple Residual MLP Framework

Data-Driven Offline Optimization for Architecting Hardware Accelerators

P-Adapters: Robustly Extracting Factual Information from Language Models with Diverse Prompts

Differentiable Scaffolding Tree for Molecule Optimization

Continuously Discovering Novel Strategies via Reward-Switching Policy Optimization

Associated Learning: an Alternative to End-to-End Backpropagation that Works on CNN, RNN, and Transformer

Maximizing Ensemble Diversity in Deep Reinforcement Learning

One After Another: Learning Incremental Skills for a Changing World

Do Users Benefit From Interpretable Vision? A User Study, Baseline, And Dataset

Revisiting flow generative models for Out-of-distribution detection

GradMax: Growing Neural Networks using Gradient Information

A global convergence theory for deep ReLU implicit networks via over-parameterization

Cross-Trajectory Representation Learning for Zero-Shot Generalization in RL

A Reduction-Based Framework for Conservative Bandits and Reinforcement Learning

Improving Non-Autoregressive Translation Models Without Distillation

Finetuned Language Models are Zero-Shot Learners

Efficient Active Search for Combinatorial Optimization Problems

Subspace Regularizers for Few-Shot Class Incremental Learning

Finite-Time Convergence and Sample Complexity of Multi-Agent Actor-Critic Reinforcement Learning with Average Reward

Planning in Stochastic Environments with a Learned Model

Graph-Guided Network for Irregularly Sampled Multivariate Time Series

LOSSY COMPRESSION WITH DISTRIBUTION SHIFT AS ENTROPY CONSTRAINED OPTIMAL TRANSPORT

POETREE: Interpretable Policy Learning with Adaptive Decision Trees

Latent Variable Sequential Set Transformers for Joint Multi-Agent Motion Prediction

Leveraging unlabeled data to predict out-of-distribution performance

Train Short, Test Long: Attention with Linear Biases Enables Input Length Extrapolation

GradSign: Model Performance Inference with Theoretical Insights

The MultiBERTs: BERT Reproductions for Robustness Analysis

On the Pitfalls of Heteroscedastic Uncertainty Estimation with Probabilistic Neural Networks

Provable Adaptation across Multiway Domains via Representation Learning

Open-Set Recognition: A Good Closed-Set Classifier is All You Need

Discrete Representations Strengthen Vision Transformer Robustness

Understanding Dimensional Collapse in Contrastive Self-supervised Learning

Likelihood Training of Schrödinger Bridge using Forward-Backward SDEs Theory

Permutation-Based SGD: Is Random Optimal?

A Fine-Tuning Approach to Belief State Modeling

Curriculum learning as a tool to uncover learning principles in the brain

Scaling Laws for Neural Machine Translation

Is Importance Weighting Incompatible with Interpolating Classifiers?

Prototype memory and attention mechanisms for few shot image generation

What Happens after SGD Reaches Zero Loss? --A Mathematical Framework

The Uncanny Similarity of Recurrence and Depth

Learning 3D Representations of Molecular Chirality with Invariance to Bond Rotations

ComPhy: Compositional Physical Reasoning of Objects and Events from Videos

A NON-PARAMETRIC REGRESSION VIEWPOINT : GENERALIZATION OF OVERPARAMETRIZED DEEP RELU NETWORK UNDER NOISY OBSERVATIONS

Independent SE(3)-Equivariant Models for End-to-End Rigid Protein Docking

An Explanation of In-context Learning as Implicit Bayesian Inference

Synchromesh: Reliable Code Generation from Pre-trained Language Models

Online Adversarial Attacks

On the Uncomputability of Partition Functions in Energy-Based Sequence Models

Improved deterministic l2 robustness on CIFAR-10 and CIFAR-100

On the Role of Neural Collapse in Transfer Learning

Blaschke Product Neural Networks (BPNN): A Physics-Infused Neural Network for Phase Retrieval of Meromorphic Functions

Diurnal or Nocturnal? Federated Learning of Multi-branch Networks from Periodically Shifting Distributions

Superclass-Conditional Gaussian Mixture Model For Learning Fine-Grained Embeddings

MetaMorph: Learning Universal Controllers with Transformers

Stiffness-aware neural network for learning Hamiltonian systems

Surrogate NAS Benchmarks: Going Beyond the Limited Search Spaces of Tabular NAS Benchmarks

Towards Training Billion Parameter Graph Neural Networks for Atomic Simulations

Bootstrapped Meta-Learning

ProtoRes: Proto-Residual Network for Pose Authoring via Learned Inverse Kinematics

BAM: Bayes with Adaptive Memory

Bundle Networks: Fiber Bundles, Local Trivializations, and a Generative Approach to Exploring Many-to-one Maps

Feature Kernel Distillation

A Biologically Interpretable Graph Convolutional Network to Link Genetic Risk Pathways and Imaging Phenotypes of Disease

Meta-Learning with Fewer Tasks through Task Interpolation

Predicting Physics in Mesh-reduced Space with Temporal Attention

Provably Robust Adversarial Examples

PER-ETD: A Polynomially Efficient Emphatic Temporal Difference Learning Method

Granger causal inference on DAGs identifies genomic loci regulating transcription

Object Pursuit: Building a Space of Objects via Discriminative Weight Generation

Shallow and Deep Networks are Near-Optimal Approximators of Korobov Functions

Cross-Lingual Transfer with Class-Weighted Language-Invariant Representations

Expressiveness and Approximation Properties of Graph Neural Networks

Almost Tight L0-norm Certified Robustness of Top-k Predictions against Adversarial Perturbations

Map Induction: Compositional spatial submap learning for efficient exploration in novel environments

GDA-AM: ON THE EFFECTIVENESS OF SOLVING MIN-IMAX OPTIMIZATION VIA ANDERSON MIXING

Evaluating Distributional Distortion in Neural Language Modeling

Triangle and Four Cycle Counting with Predictions in Graph Streams

Trivial or Impossible --- dichotomous data difficulty masks model differences (on ImageNet and beyond)

Dual Lottery Ticket Hypothesis

Selective Ensembles for Consistent Predictions

(ends 7:30 PM)

7 p.m.

Affinity Social:

Queer in AI Social @ ICLR 2022

(duration 1.0 hr)

8 p.m.

10 p.m.

WED 27 APR

midnight

Invited Talk:

‘Affordances’ for Machine Learning

Jenny Davis

(ends 1:15 AM)

1:30 a.m.

Poster Session 6 [1:30-3:30]

Fine-Tuning can Distort Pretrained Features and Underperform Out-of-Distribution

TAPEX: Table Pre-training via Learning a Neural SQL Executor

Causal Contextual Bandits with Targeted Interventions

Recycling Model Updates in Federated Learning: Are Gradient Subspaces Low-Rank?

Multi-Mode Deep Matrix and Tensor Factorization

Representing Mixtures of Word Embeddings with Mixtures of Topic Embeddings

Sampling with Mirrored Stein Operators

Towards Understanding the Data Dependency of Mixup-style Training

Learning to Remember Patterns: Pattern Matching Memory Networks for Traffic Forecasting

Actor-Critic Policy Optimization in a Large-Scale Imperfect-Information Game

Learning Towards The Largest Margins

X-model: Improving Data Efficiency in Deep Learning with A Minimax Model

Transformer Embeddings of Irregularly Spaced Events and Their Participants

Conditional Contrastive Learning with Kernel

Training Transition Policies via Distribution Matching for Complex Tasks

Procedural generalization by planning with self-supervised world models

Backdoor Defense via Decoupling the Training Process

Multi-objective Optimization by Learning Space Partition

Skill-based Meta-Reinforcement Learning

Differentially Private Fine-tuning of Language Models

Does your graph need a confidence boost? Convergent boosted smoothing on graphs with tabular node features

Graph-Enhanced Exploration for Goal-oriented Reinforcement Learning

Understanding Latent Correlation-Based Multiview Learning and Self-Supervision: An Identifiability Perspective

Towards Deployment-Efficient Reinforcement Learning: Lower Bound and Optimality

NASI: Label- and Data-agnostic Neural Architecture Search at Initialization

GLASS: GNN with Labeling Tricks for Subgraph Representation Learning

Latent Image Animator: Learning to Animate Images via Latent Space Navigation

TAda! Temporally-Adaptive Convolutions for Video Understanding

Learning to Complete Code with Sketches

Robust Unlearnable Examples: Protecting Data Privacy Against Adversarial Learning

Equivariant Graph Mechanics Networks with Constraints

Invariant Causal Representation Learning for Out-of-Distribution Generalization

Adaptive Wavelet Transformer Network for 3D Shape Representation Learning

An Autoregressive Flow Model for 3D Molecular Geometry Generation from Scratch

Maximum n-times Coverage for Vaccine Design

Recursive Disentanglement Network

Bandit Learning with Joint Effect of Incentivized Sampling, Delayed Sampling Feedback, and Self-Reinforcing User Preferences

When Can We Learn General-Sum Markov Games with a Large Number of Players Sample-Efficiently?

Pareto Policy Pool for Model-based Offline Reinforcement Learning

VOS: Learning What You Don't Know by Virtual Outlier Synthesis

SURF: Semi-supervised Reward Learning with Data Augmentation for Feedback-efficient Preference-based Reinforcement Learning

Open-World Semi-Supervised Learning

Why Propagate Alone? Parallel Use of Labels and Features on Graphs

Neural Structured Prediction for Inductive Node Classification

Dynamics-Aware Comparison of Learned Reward Functions

Learning Features with Parameter-Free Layers

PoNet: Pooling Network for Efficient Token Mixing in Long Sequences

On feature learning in neural networks with global convergence guarantees

A Unified Wasserstein Distributional Robustness Framework for Adversarial Training

A Theoretical Analysis on Feature Learning in Neural Networks: Emergence from Inputs and Advantage over Fixed Features

Learning to Annotate Part Segmentation with Gradient Matching

Strength of Minibatch Noise in SGD

Optimal Representations for Covariate Shift

Discovering Invariant Rationales for Graph Neural Networks

Learning Pruning-Friendly Networks via Frank-Wolfe: One-Shot, Any-Sparsity, And No Retraining

Knowledge Removal in Sampling-based Bayesian Inference

SDEdit: Guided Image Synthesis and Editing with Stochastic Differential Equations

CoordX: Accelerating Implicit Neural Representation with a Split MLP Architecture

Fixed Neural Network Steganography: Train the images, not the network

Inductive Relation Prediction Using Analogy Subgraph Embeddings

Accelerated Policy Learning with Parallel Differentiable Simulation

Finding Biological Plausibility for Adversarially Robust Features via Metameric Tasks

Sparse Attention with Learning to Hash

Is Homophily a Necessity for Graph Neural Networks?

DKM: Differentiable k-Means Clustering Layer for Neural Network Compression

Creating Training Sets via Weak Indirect Supervision

Improving Federated Learning Face Recognition via Privacy-Agnostic Clusters

Neural Deep Equilibrium Solvers

SimVLM: Simple Visual Language Model Pretraining with Weak Supervision

A Class of Short-term Recurrence Anderson Mixing Methods and Their Applications

How Well Does Self-Supervised Pre-Training Perform with Streaming Data?

The Rich Get Richer: Disparate Impact of Semi-Supervised Learning

Towards General Function Approximation in Zero-Sum Markov Games

Back2Future: Leveraging Backfill Dynamics for Improving Real-time Predictions in Future

Should We Be Pre-training? An Argument for End-task Aware Training as an Alternative

How Does SimSiam Avoid Collapse Without Negative Samples? A Unified Understanding with Self-supervised Contrastive Learning

Disentanglement Analysis with Partial Information Decomposition

Bi-linear Value Networks for Multi-goal Reinforcement Learning

Towards Understanding Generalization via Decomposing Excess Risk Dynamics

Neural Stochastic Dual Dynamic Programming

(ends 3:30 AM)

8 a.m.

Invited Talk:

Do you see what I see? Large-scale learning from multimodal videos

Cordelia Schmid

(ends 9:15 AM)

9:30 a.m.

Poster Session 7 [9:30-11:30]

Deep Attentive Variational Inference

The Neural Data Router: Adaptive Control Flow in Transformers Improves Systematic Generalization

SHINE: SHaring the INverse Estimate from the forward pass for bi-level optimization and implicit models

Mapping conditional distributions for domain adaptation under generalized target shift

Normalization of Language Embeddings for Cross-Lingual Alignment

Filling the G_ap_s: Multivariate Time Series Imputation by Graph Neural Networks

Rethinking Goal-Conditioned Supervised Learning and Its Connection to Offline RL

On the Optimal Memorization Power of ReLU Neural Networks

On Redundancy and Diversity in Cell-based Neural Architecture Search

Complete Verification via Multi-Neuron Relaxation Guided Branch-and-Bound

Model Agnostic Interpretability for Multiple Instance Learning

Graph Neural Networks with Learnable Structural and Positional Representations

NETWORK INSENSITIVITY TO PARAMETER NOISE VIA PARAMETER ATTACK DURING TRAINING

Online Facility Location with Predictions

When, Why, and Which Pretrained GANs Are Useful?

On the Importance of Difficulty Calibration in Membership Inference Attacks

Multi-Agent MDP Homomorphic Networks

CycleMLP: A MLP-like Architecture for Dense Prediction

When should agents explore?

RotoGrad: Gradient Homogenization in Multitask Learning

Learning a subspace of policies for online adaptation in Reinforcement Learning

Visual hyperacuity with moving sensor and recurrent neural computations

Graphon based Clustering and Testing of Networks: Algorithms and Theory

Fast Differentiable Matrix Square Root

Online Continual Learning on Class Incremental Blurry Task Configuration with Anytime Inference

On the Limitations of Multimodal VAEs

Natural Posterior Network: Deep Bayesian Predictive Uncertainty for Exponential Family Distributions

Neural Link Prediction with Walk Pooling

Pessimistic Bootstrapping for Uncertainty-Driven Offline Reinforcement Learning

Discrepancy-Based Active Learning for Domain Adaptation

NASPY: Automated Extraction of Automated Machine Learning Models

Learnability of convolutional neural networks for infinite dimensional input via mixed and anisotropic smoothness

Minibatch vs Local SGD with Shuffling: Tight Convergence Bounds and Beyond

LIGS: Learnable Intrinsic-Reward Generation Selection for Multi-Agent Learning

GATSBI: Generative Adversarial Training for Simulation-Based Inference

Generalization of Neural Combinatorial Solvers Through the Lens of Adversarial Robustness

cosFormer: Rethinking Softmax In Attention

On Distributed Adaptive Optimization with Gradient Compression

Switch to Generalize: Domain-Switch Learning for Cross-Domain Few-Shot Classification

iFlood: A Stable and Effective Regularizer

Discovering Latent Concepts Learned in BERT

Learning Multimodal VAEs through Mutual Supervision

Conditioning Sequence-to-sequence Networks with Learned Activations

Implicit Bias of Adversarial Training for Deep Neural Networks

Proving the Lottery Ticket Hypothesis for Convolutional Neural Networks

PolyLoss: A Polynomial Expansion Perspective of Classification Loss Functions

Case-based reasoning for better generalization in textual reinforcement learning

HyAR: Addressing Discrete-Continuous Action Reinforcement Learning via Hybrid Action Representation

Optimal Transport for Causal Discovery

Understanding and Preventing Capacity Loss in Reinforcement Learning

SPIRAL: Self-supervised Perturbation-Invariant Representation Learning for Speech Pre-Training

Neural Network Approximation based on Hausdorff distance of Tropical Zonotopes

A fast and accurate splitting method for optimal transport: analysis and implementation

FP-DETR: Detection Transformer Advanced by Fully Pre-training

Attacking deep networks with surrogate-based adversarial black-box methods is easy

Neural graphical modelling in continuous-time: consistency guarantees and algorithms

On-Policy Model Errors in Reinforcement Learning

What’s Wrong with Deep Learning in Tree Search for Combinatorial Optimization

A Johnson-Lindenstrauss Framework for Randomly Initialized CNNs

Evaluating Model-Based Planning and Planner Amortization for Continuous Control

Anomaly Detection for Tabular Data with Internal Contrastive Learning

Frame Averaging for Invariant and Equivariant Network Design

Wiring Up Vision: Minimizing Supervised Synaptic Updates Needed to Produce a Primate Ventral Stream

WeakM3D: Towards Weakly Supervised Monocular 3D Object Detection

Spike-inspired rank coding for fast and accurate recurrent neural networks

Learning Super-Features for Image Retrieval

Multi-Task Processes

Distributionally Robust Fair Principal Components via Geodesic Descents

GRAND++: Graph Neural Diffusion with A Source Term

Generalization Through the Lens of Leave-One-Out Error

Learning transferable motor skills with hierarchical latent mixture policies

How to Train Your MAML to Excel in Few-Shot Classification

MonoDistill: Learning Spatial Features for Monocular 3D Object Detection

(ends 11:30 AM)

11 a.m.

4 p.m.

Oral 3: Learning from distribution shift [4:00-5:30]

Oral s 4:00-5:30

[4:00] Fine-Tuning can Distort Pretrained Features and Underperform Out-of-Distribution

[4:15] Asymmetry Learning for Counterfactually-invariant Classification in OOD Tasks

[4:30] A Fine-Grained Analysis on Distribution Shift

[4:45] Sparse Communication via Mixed Distributions

[5:00] Frame Averaging for Invariant and Equivariant Network Design

[5:15] F8Net: Fixed-Point 8-bit Only Multiplication for Network Quantization

(ends 5:30 PM)

Oral 3: Meta-learning and adaptation [4:00-5:30]

Oral s 4:00-5:30

[4:00] Bootstrapped Meta-Learning

[4:15] Coordination Among Neural Modules Through a Shared Global Workspace

[4:30] Meta-Learning with Fewer Tasks through Task Interpolation

[4:45] Weighted Training for Cross-Task Learning

[5:00] Domino: Discovering Systematic Errors with Cross-Modal Embeddings

[5:15] Extending the WILDS Benchmark for Unsupervised Adaptation

(ends 5:30 PM)

5:30 p.m.

Poster Session 8 [5:30-7:30]

ARTEMIS: Attention-based Retrieval with Text-Explicit Matching and Implicit Similarity

Heteroscedastic Temporal Variational Autoencoder For Irregularly Sampled Time Series

A Comparison of Hamming Errors of Representative Variable Selection Methods

Neural Methods for Logical Reasoning over Knowledge Graphs

Auto-scaling Vision Transformers without Training

Topologically Regularized Data Embeddings

Assessing Generalization of SGD via Disagreement

Variational autoencoders in the presence of low-dimensional data: landscape and implicit bias

Supervision Exists Everywhere: A Data Efficient Contrastive Language-Image Pre-training Paradigm

Learning Fast Samplers for Diffusion Models by Differentiating Through Sample Quality

Adversarial Unlearning of Backdoors via Implicit Hypergradient

How Attentive are Graph Attention Networks?

Linking Emergent and Natural Languages via Corpus Transfer

Multimeasurement Generative Models

Generative Planning for Temporally Coordinated Exploration in Reinforcement Learning

Graph-Relational Domain Adaptation

Value Function Spaces: Skill-Centric State Abstractions for Long-Horizon Reasoning

Memorizing Transformers

Pseudo Numerical Methods for Diffusion Models on Manifolds

Capturing Structural Locality in Non-parametric Language Models

Learning Prototype-oriented Set Representations for Meta-Learning

Sample Efficient Deep Reinforcement Learning via Uncertainty Estimation

Implicit Bias of Projected Subgradient Method Gives Provable Robust Recovery of Subspaces of Unknown Codimension

SphereFace2: Binary Classification is All You Need for Deep Face Recognition

Label Leakage and Protection in Two-party Split Learning

Fortuitous Forgetting in Connectionist Networks

PAC-Bayes Information Bottleneck

Bayesian Framework for Gradient Leakage

Plant 'n' Seek: Can You Find the Winning Ticket?

Large Learning Rate Tames Homogeneity: Convergence and Balancing Effect

Differentiable DAG Sampling

Anomaly Transformer: Time Series Anomaly Detection with Association Discrepancy

Knowledge Infused Decoding

The Information Geometry of Unsupervised Reinforcement Learning

Learnability Lock: Authorized Learnability Control Through Adversarial Invertible Transformations

CrossMatch: Cross-Classifier Consistency Regularization for Open-Set Single Domain Generalization

Offline Reinforcement Learning with Value-based Episodic Memory

Towards Understanding the Robustness Against Evasion Attack on Categorical Data

Reward Uncertainty for Exploration in Preference-based Reinforcement Learning

Cross-Domain Imitation Learning via Optimal Transport

Poisoning and Backdooring Contrastive Learning

Space-Time Graph Neural Networks

On the Certified Robustness for Ensemble Models and Beyond

Gaussian Mixture Convolution Networks

Discovering Nonlinear PDEs from Scarce Data with Physics-encoded Learning

Provably Filtering Exogenous Distractors using Multistep Inverse Dynamics

CLEVA-Compass: A Continual Learning Evaluation Assessment Compass to Promote Research Transparency and Comparability

Phenomenology of Double Descent in Finite-Width Neural Networks

Transformer-based Transform Coding

Orchestrated Value Mapping for Reinforcement Learning

Information Prioritization through Empowerment in Visual Model-based RL

GeneDisco: A Benchmark for Experimental Design in Drug Discovery

Constructing a Good Behavior Basis for Transfer using Generalized Policy Updates

Better Supervisory Signals by Observing Learning Paths

Exposing the Implicit Energy Networks behind Masked Language Models via Metropolis--Hastings

Learning Distributionally Robust Models at Scale via Composite Optimization

No One Representation to Rule Them All: Overlapping Features of Training Methods

Value Gradient weighted Model-Based Reinforcement Learning

Data Poisoning Won’t Save You From Facial Recognition

Amortized Tree Generation for Bottom-up Synthesis Planning and Synthesizable Molecular Design

Task-Induced Representation Learning

HyperDQN: A Randomized Exploration Method for Deep Reinforcement Learning

Understanding and Leveraging Overparameterization in Recursive Value Estimation

Fast Regression for Structured Inputs

Practical Conditional Neural Process Via Tractable Dependent Predictions

On the Importance of Firth Bias Reduction in Few-Shot Classification

DARA: Dynamics-Aware Reward Augmentation in Offline Reinforcement Learning

Contact Points Discovery for Soft-Body Manipulations with Differentiable Physics

Generalized rectifier wavelet covariance models for texture synthesis

Deep Learning without Shortcuts: Shaping the Kernel with Tailored Rectifiers

ExT5: Towards Extreme Multi-Task Scaling for Transfer Learning

FILIP: Fine-grained Interactive Language-Image Pre-Training

Comparing Distributions by Measuring Differences that Affect Decision Making

Understanding over-squashing and bottlenecks on graphs via curvature

The Effects of Invertibility on the Representational Complexity of Encoders in Variational Autoencoders

Implicit Bias of MSE Gradient Optimization in Underparameterized Neural Networks

Hindsight Foresight Relabeling for Meta-Reinforcement Learning

Missingness Bias in Model Debugging

Relational Multi-Task Learning: Modeling Relations between Data and Tasks

Hidden Parameter Recurrent State Space Models For Changing Dynamics Scenarios

Non-Transferable Learning: A New Approach for Model Ownership Verification and Applicability Authorization

Scene Transformer: A unified architecture for predicting future trajectories of multiple agents

Learning to Dequantise with Truncated Flows

From Intervention to Domain Transportation: A Novel Perspective to Optimize Recommendation

Memory Augmented Optimizers for Deep Learning

RvS: What is Essential for Offline RL via Supervised Learning?

A Statistical Framework for Efficient Out of Distribution Detection in Deep Neural Networks

Step-unrolled Denoising Autoencoders for Text Generation

DR3: Value-Based Deep Reinforcement Learning Requires Explicit Regularization

CoMPS: Continual Meta Policy Search

The Boltzmann Policy Distribution: Accounting for Systematic Suboptimality in Human Models

Decentralized Learning for Overparameterized Problems: A Multi-Agent Kernel Approximation Approach

Learning State Representations via Retracing in Reinforcement Learning

Graph Auto-Encoder via Neighborhood Wasserstein Reconstruction

Churn Reduction via Distillation

On the role of population heterogeneity in emergent communication

AdaMatch: A Unified Approach to Semi-Supervised Learning and Domain Adaptation

EE-Net: Exploitation-Exploration Neural Networks in Contextual Bandits

Consistent Counterfactuals for Deep Models

Connectome-constrained Latent Variable Model of Whole-Brain Neural Activity

Towards Building A Group-based Unsupervised Representation Disentanglement Framework

DIVA: Dataset Derivative of a Learning Task

$\pi$BO: Augmenting Acquisition Functions with User Beliefs for Bayesian Optimization

Distribution Compression in Near-Linear Time

Long Expressive Memory for Sequence Modeling

Efficient Token Mixing for Transformers via Adaptive Fourier Neural Operators

Image BERT Pre-training with Online Tokenizer

Pessimistic Model-based Offline Reinforcement Learning under Partial Coverage

Diverse Client Selection for Federated Learning via Submodular Maximization

On Improving Adversarial Transferability of Vision Transformers

The Convex Geometry of Backpropagation: Neural Network Gradient Flows Converge to Extreme Points of the Dual Convex Program

The Hidden Convex Optimization Landscape of Regularized Two-Layer ReLU Networks: an Exact Characterization of Optimal Solutions

CrowdPlay: Crowdsourcing Human Demonstrations for Offline Learning

Sequence Approximation using Feedforward Spiking Neural Network for Spatiotemporal Learning: Theory and Optimization Methods

Learning Representation from Neural Fisher Kernel with Low-rank Approximation

The Efficiency Misnomer

(ends 7:30 PM)

7 p.m.

THU 28 APR

midnight

Invited Talk:

Accelerating AI Systems: Let the Data Flow!

Kunle Olukotun

(ends 1:15 AM)

1:30 a.m.

Poster Session 9 [1:30-3:30]

Know Thyself: Transferable Visual Control Policies Through Robot-Awareness

How to Robustify Black-Box ML Models? A Zeroth-Order Optimization Perspective

Bridging Recommendation and Marketing via Recurrent Intensity Modeling

Online Hyperparameter Meta-Learning with Hypergradient Distillation

EViT: Expediting Vision Transformers via Token Reorganizations

Node Feature Extraction by Self-Supervised Multi-scale Neighborhood Prediction

RelViT: Concept-guided Vision Transformer for Visual Relational Reasoning

LEARNING GUARANTEES FOR GRAPH CONVOLUTIONAL NETWORKS ON THE STOCHASTIC BLOCK MODEL

Learning Weakly-supervised Contrastive Representations

AdaAug: Learning Class- and Instance-adaptive Data Augmentation Policies

Stein Latent Optimization for Generative Adversarial Networks

Evading Adversarial Example Detection Defenses with Orthogonal Projected Gradient Descent

QDrop: Randomly Dropping Quantization for Extremely Low-bit Post-Training Quantization

Mention Memory: incorporating textual knowledge into Transformers through entity mention attention

Particle Stochastic Dual Coordinate Ascent: Exponential convergent algorithm for mean field neural network optimization

Learned Simulators for Turbulence

C-Planning: An Automatic Curriculum for Learning Goal-Reaching Tasks

SGD Can Converge to Local Maxima

Large Language Models Can Be Strong Differentially Private Learners

Pre-training Molecular Graph Representation with 3D Geometry

PipeGCN: Efficient Full-Graph Training of Graph Convolutional Networks with Pipelined Feature Communication

Patch-Fool: Are Vision Transformers Always Robust Against Adversarial Perturbations?

Path Auxiliary Proposal for MCMC in Discrete Space

Neural Solvers for Fast and Accurate Numerical Optimal Control

Optimization inspired Multi-Branch Equilibrium Models

Meta Learning Low Rank Covariance Factors for Energy Based Deterministic Uncertainty

Is Fairness Only Metric Deep? Evaluating and Addressing Subgroup Gaps in Deep Metric Learning

A Theory of Tournament Representations

Learning with Noisy Labels Revisited: A Study Using Real-World Human Annotations

Pseudo-Labeled Auto-Curriculum Learning for Semi-Supervised Keypoint Localization

Pretrained Language Model in Continual Learning: A Comparative Study

Contrastive Clustering to Mine Pseudo Parallel Data for Unsupervised Translation

Convergent and Efficient Deep Q Learning Algorithm

On the Convergence of the Monte Carlo Exploring Starts Algorithm for Reinforcement Learning

Concurrent Adversarial Learning for Large-Batch Training

Symbolic Learning to Optimize: Towards Interpretability and Scalability

Eigencurve: Optimal Learning Rate Schedule for SGD on Quadratic Objectives with Skewed Hessian Spectrums

Meta-Imitation Learning by Watching Video Demonstrations

Context-Aware Sparse Deep Coordination Graphs

Scale Efficiently: Insights from Pretraining and Finetuning Transformers

Cold Brew: Distilling Graph Node Representations with Incomplete or Missing Neighborhoods

Benchmarking the Spectrum of Agent Capabilities

Declarative nets that are equilibrium models

BEiT: BERT Pre-Training of Image Transformers

Quantitative Performance Assessment of CNN Units via Topological Entropy Calculation

TRGP: Trust Region Gradient Projection for Continual Learning

Generalizing Few-Shot NAS with Gradient Matching

Learning Continuous Environment Fields via Implicit Functions

On the Convergence of Certified Robust Training with Interval Bound Propagation

DEPTS: Deep Expansion Learning for Periodic Time Series Forecasting

Half-Inverse Gradients for Physical Deep Learning

F8Net: Fixed-Point 8-bit Only Multiplication for Network Quantization

Learning Versatile Neural Architectures by Propagating Network Codes

Efficient Sharpness-aware Minimization for Improved Training of Neural Networks

Convergent Graph Solvers

Responsible Disclosure of Generative Models Using Scalable Fingerprinting

Zero-Shot Self-Supervised Learning for MRI Reconstruction

Neural Networks as Kernel Learners: The Silent Alignment Effect

The Effects of Reward Misspecification: Mapping and Mitigating Misaligned Models

Coherence-based Label Propagation over Time Series for Accelerated Active Learning

Clean Images are Hard to Reblur: Exploiting the Ill-Posed Inverse Task for Dynamic Scene Deblurring

Neural Collapse Under MSE Loss: Proximity to and Dynamics on the Central Path

T-WaveNet: A Tree-Structured Wavelet Neural Network for Time Series Signal Analysis

Gradient Information Matters in Policy Optimization by Back-propagating through Model

Certified Robustness for Deep Equilibrium Models via Interval Bound Propagation

Exploring extreme parameter compression for pre-trained language models

CDTrans: Cross-domain Transformer for Unsupervised Domain Adaptation

Tracking the risk of a deployed model and detecting harmful distribution shifts

Information Gain Propagation: a New Way to Graph Active Learning with Soft Labels

Efficient Learning of Safe Driving Policy via Human-AI Copilot Optimization

Omni-Scale CNNs: a simple and effective kernel size configuration for time series classification

How to Inject Backdoors with Better Consistency: Logit Anchoring on Clean Data

NODE-GAM: Neural Generalized Additive Model for Interpretable Deep Learning

Beyond ImageNet Attack: Towards Crafting Adversarial Examples for Black-box Domains

How Low Can We Go: Trading Memory for Error in Low-Precision Training

Charformer: Fast Character Transformers via Gradient-based Subword Tokenization

$\beta$-Intact-VAE: Identifying and Estimating Causal Effects under Limited Overlap

Toward Efficient Low-Precision Training: Data Format Optimization and Hysteresis Quantization

(ends 3:30 AM)

3 a.m.

8 a.m.

Oral 4: Sequence modeling [8:00-9:30]

Oral s 8:00-9:30

[8:00] Diffusion-Based Voice Conversion with Fast Maximum Likelihood Sampling Scheme

[8:15] Natural Language Descriptions of Deep Features

[8:30] Finetuned Language Models are Zero-Shot Learners

[8:45] Large Language Models Can Be Strong Differentially Private Learners

[9:00] GeoDiff: A Geometric Diffusion Model for Molecular Conformation Generation

[9:15] Pyraformer: Low-Complexity Pyramidal Attention for Long-Range Time Series Modeling and Forecasting

(ends 9:30 AM)

Oral 4: Probablistic Models, Vision [8:00-9:30]

Oral s 8:00-9:30

[8:00] Analytic-DPM: an Analytic Estimate of the Optimal Reverse Variance in Diffusion Probabilistic Models

[8:15] Comparing Distributions by Measuring Differences that Affect Decision Making

[8:30] Unsupervised Vision-Language Grammar Induction with Shared Structure Modeling

[8:45] RISP: Rendering-Invariant State Predictor with Differentiable Simulation and Rendering for Cross-Domain Parameter Estimation

[9:00] BEiT: BERT Pre-Training of Image Transformers

[9:15] Resolving Training Biases via Influence-based Data Relabeling

(ends 9:30 AM)

9:30 a.m.

Poster Session 10 [9:30-11:30]

CURVATURE-GUIDED DYNAMIC SCALE NETWORKS FOR MULTI-VIEW STEREO

Learning to Extend Molecular Scaffolds with Structural Motifs

FlexConv: Continuous Kernel Convolutions With Differentiable Kernel Sizes

Unraveling Model-Agnostic Meta-Learning via The Adaptation Learning Rate

DISCOVERING AND EXPLAINING THE REPRESENTATION BOTTLENECK OF DNNS

Policy improvement by planning with Gumbel

Deep Point Cloud Reconstruction

How Do Vision Transformers Work?

GNN-LM: Language Modeling based on Global Contexts via GNN

A Deep Variational Approach to Clustering Survival Data

Distilling GANs with Style-Mixed Triplets for X2I Translation with Limited Data

Learning Causal Models from Conditional Moment Restrictions by Importance Weighting

ToM2C: Target-oriented Multi-agent Communication and Cooperation with Theory of Mind

Federated Learning from Only Unlabeled Data with Class-conditional-sharing Clients

Reinforcement Learning in Presence of Discrete Markovian Context Evolution

Nonlinear ICA Using Volume-Preserving Transformations

Uncertainty Modeling for Out-of-Distribution Generalization

Ab-Initio Potential Energy Surfaces by Pairing GNNs with Neural Wave Functions

Group equivariant neural posterior estimation

AlphaZero-based Proof Cost Network to Aid Game Solving

Embedded-model flows: Combining the inductive biases of model-free deep learning and explicit probabilistic modeling

Generative Pseudo-Inverse Memory

Sparsity Winning Twice: Better Robust Generalization from More Efficient Training

Pareto Set Learning for Neural Multi-Objective Combinatorial Optimization

CADDA: Class-wise Automatic Differentiable Data Augmentation for EEG Signals

Variational Predictive Routing with Nested Subjective Timescales

ADAVI: Automatic Dual Amortized Variational Inference Applied To Pyramidal Bayesian Models

Evaluating Disentanglement of Structured Representations

Exploiting Class Activation Value for Partial-Label Learning

Omni-Dimensional Dynamic Convolution

Scalable One-Pass Optimisation of High-Dimensional Weight-Update Hyperparameters by Implicit Differentiation

SOSP: Efficiently Capturing Global Correlations by Second-Order Structured Pruning

Pyraformer: Low-Complexity Pyramidal Attention for Long-Range Time Series Modeling and Forecasting

Dynamic Token Normalization improves Vision Transformers

CrossFormer: A Versatile Vision Transformer Hinging on Cross-scale Attention

Prototypical Contrastive Predictive Coding

Dive Deeper Into Integral Pose Regression

Partial Wasserstein Adversarial Network for Non-rigid Point Set Registration

CKConv: Continuous Kernel Convolution For Sequential Data

Understanding approximate and unrolled dictionary learning for pattern recovery

On the Pitfalls of Analyzing Individual Neurons in Language Models

IGLU: Efficient GCN Training via Lazy Updates

LORD: Lower-Dimensional Embedding of Log-Signature in Neural Rough Differential Equations

CoBERL: Contrastive BERT for Reinforcement Learning

Differentiable Expectation-Maximization for Set Representation Learning

PEARL: Data Synthesis via Private Embeddings and Adversarial Reconstruction Learning

Vitruvion: A Generative Model of Parametric CAD Sketches

Near-Optimal Reward-Free Exploration for Linear Mixture MDPs with Plug-in Solver

Explaining Point Processes by Learning Interpretable Temporal Logic Rules

On Covariate Shift of Latent Confounders in Imitation and Reinforcement Learning

Learning Discrete Structured Variational Auto-Encoder using Natural Evolution Strategies

Universal Approximation Under Constraints is Possible with Transformers

D-CODE: Discovering Closed-form ODEs from Observed Trajectories

Direct then Diffuse: Incremental Unsupervised Skill Discovery for State Covering and Goal Reaching

Self-supervised Learning is More Robust to Dataset Imbalance

Evaluation Metrics for Graph Generative Models: Problems, Pitfalls, and Practical Solutions

FedPara: Low-rank Hadamard Product for Communication-Efficient Federated Learning

Counterfactual Plans under Distributional Ambiguity

Active Hierarchical Exploration with Stable Subgoal Representation Learning

Spread Spurious Attribute: Improving Worst-group Accuracy with Spurious Attribute Estimation

Robust and Scalable SDE Learning: A Functional Perspective

Multi-Critic Actor Learning: Teaching RL Policies to Act with Style

Adversarially Robust Conformal Prediction

Crystal Diffusion Variational Autoencoder for Periodic Material Generation

Post hoc Explanations may be Ineffective for Detecting Unknown Spurious Correlation

Constraining Linear-chain CRFs to Regular Languages

Entroformer: A Transformer-based Entropy Model for Learned Image Compression

Representation-Agnostic Shape Fields

High Probability Bounds for a Class of Nonconvex Algorithms with AdaGrad Stepsize

Defending Against Image Corruptions Through Adversarial Augmentations

StyleNeRF: A Style-based 3D Aware Generator for High-resolution Image Synthesis

Exploring Memorization in Adversarial Training

Sequential Reptile: Inter-Task Gradient Alignment for Multilingual Learning

AS-MLP: An Axial Shifted MLP Architecture for Vision

PSA-GAN: Progressive Self Attention GANs for Synthetic Time Series

It Takes Two to Tango: Mixup for Deep Metric Learning

Training Data Generating Networks: Shape Reconstruction via Bi-level Optimization

Neural Markov Controlled SDE: Stochastic Optimization for Continuous-Time Data

Reducing Excessive Margin to Achieve a Better Accuracy vs. Robustness Trade-off

You Mostly Walk Alone: Analyzing Feature Attribution in Trajectory Prediction

(ends 11:30 AM)

11 a.m.

1 p.m.

4 p.m.

From Reinforcement Learning to AI:

Doina Precup

(ends 5:15 PM)

5:30 p.m.

Poster Session 11 [5:30-7:30]

Message Passing Neural PDE Solvers

Unsupervised Learning of Full-Waveform Inversion: Connecting CNN and Partial Differential Equation in a Loop

Learning Curves for Gaussian Process Regression with Power-Law Priors and Targets

Efficient Split-Mix Federated Learning for On-Demand and In-Situ Customization

Multi-Stage Episodic Control for Strategic Exploration in Text Games

Bag of Instances Aggregation Boosts Self-supervised Distillation

Geometric and Physical Quantities improve E(3) Equivariant Message Passing

Fine-grained Differentiable Physics: A Yarn-level Model for Fabrics

Non-Linear Operator Approximations for Initial Value Problems

Energy-Inspired Molecular Conformation Optimization

TRAIL: Near-Optimal Imitation Learning with Suboptimal Data

IFR-Explore: Learning Inter-object Functional Relationships in 3D Indoor Scenes

Sparse Communication via Mixed Distributions

Imitation Learning from Observations under Transition Model Disparity

Hindsight: Posterior-guided training of retrievers for improved open-ended generation

Learning Object-Oriented Dynamics for Planning from Text

Attention-based Interpretability with Concept Transformers

Asymmetry Learning for Counterfactually-invariant Classification in OOD Tasks

A Loss Curvature Perspective on Training Instabilities of Deep Learning Models

Monotonic Differentiable Sorting Networks

Learning Curves for SGD on Structured Features

Domain Adversarial Training: A Game Perspective

Differentially Private Fractional Frequency Moments Estimation with Polylogarithmic Space

DriPP: Driven Point Processes to Model Stimuli Induced Patterns in M/EEG Signals

How Did the Model Change? Efficiently Assessing Machine Learning API Shifts

Online Ad Hoc Teamwork under Partial Observability

Compositional Attention: Disentangling Search and Retrieval

Boosting Randomized Smoothing with Variance Reduced Classifiers

Representation Learning for Online and Offline RL in Low-rank MDPs

On Bridging Generic and Personalized Federated Learning for Image Classification

Learning Altruistic Behaviours in Reinforcement Learning without External Rewards

Bregman Gradient Policy Optimization

Maximum Entropy RL (Provably) Solves Some Robust RL Problems

Looking Back on Learned Experiences For Class/task Incremental Learning

An Agnostic Approach to Federated Learning with Class Imbalance

Axiomatic Explanations for Visual Search, Retrieval, and Similarity Learning

Equivariant Self-Supervised Learning: Encouraging Equivariance in Representations

Local Feature Swapping for Generalization in Reinforcement Learning

Equivariant Subgraph Aggregation Networks

Learning-Augmented $k$-means Clustering

Learning to Map for Active Semantic Goal Navigation

Query Embedding on Hyper-Relational Knowledge Graphs

DiffSkill: Skill Abstraction from Differentiable Physics for Deformable Object Manipulations with Tools

Information Bottleneck: Exact Analysis of (Quantized) Neural Networks

Pix2seq: A Language Modeling Framework for Object Detection

Hidden Convexity of Wasserstein GANs: Interpretable Generative Models with Closed-Form Solutions

GiraffeDet: A Heavy-Neck Paradigm for Object Detection

Unsupervised Semantic Segmentation by Distilling Feature Correspondences

Language model compression with weighted low-rank factorization

Communication-Efficient Actor-Critic Methods for Homogeneous Markov Games

Minimax Optimization with Smooth Algorithmic Adversaries

ClimateGAN: Raising Climate Change Awareness by Generating Images of Floods

On the Convergence of mSGD and AdaGrad for Stochastic Optimization

Divisive Feature Normalization Improves Image Recognition Performance in AlexNet

CodeTrek: Flexible Modeling of Code using an Extensible Relational Representation

Mastering Visual Continuous Control: Improved Data-Augmented Reinforcement Learning

A Zest of LIME: Towards Architecture-Independent Model Distances

MIDI-DDSP: Detailed Control of Musical Performance via Hierarchical Modeling

Low-Budget Active Learning via Wasserstein Distance: An Integer Programming Approach

Pareto Policy Adaptation

Unified Visual Transformer Compression

Can an Image Classifier Suffice For Action Recognition?

On the relation between statistical learning and perceptual distances

Explanations of Black-Box Models based on Directional Feature Interactions

An Operator Theoretic View On Pruning Deep Neural Networks

Capacity of Group-invariant Linear Readouts from Equivariant Representations: How Many Objects can be Linearly Classified Under All Possible Views?

Domino: Discovering Systematic Errors with Cross-Modal Embeddings

Optimization and Adaptive Generalization of Three layer Neural Networks

VICReg: Variance-Invariance-Covariance Regularization for Self-Supervised Learning

Diffusion-Based Voice Conversion with Fast Maximum Likelihood Sampling Scheme

Graph-less Neural Networks: Teaching Old MLPs New Tricks Via Distillation

Demystifying Limited Adversarial Transferability in Automatic Speech Recognition Systems

Learning more skills through optimistic exploration

Conditional Object-Centric Learning from Video

Autonomous Reinforcement Learning: Formalism and Benchmarking

Mirror Descent Policy Optimization

Understanding the Variance Collapse of SVGD in High Dimensions

FedChain: Chained Algorithms for Near-optimal Communication Cost in Federated Learning

Learning to Schedule Learning rate with Graph Neural Networks

Fairness in Representation for Multilingual NLP: Insights from Controlled Experiments on Conditional Language Modeling

Unifying Likelihood-free Inference with Black-box Optimization and Beyond

Anti-Concentrated Confidence Bonuses For Scalable Exploration

How many degrees of freedom do we need to train deep networks: a loss landscape perspective

EntQA: Entity Linking as Question Answering

Sample Selection with Uncertainty of Losses for Learning with Noisy Labels

Towards a Unified View of Parameter-Efficient Transfer Learning

Properties from mechanisms: an equivariance perspective on identifiable representation learning

Revisit Kernel Pruning with Lottery Regulated Grouped Convolutions

Generalized Natural Gradient Flows in Hidden Convex-Concave Games and GANs

FILM: Following Instructions in Language with Modular Methods

R4D: Utilizing Reference Objects for Long-Range Distance Estimation

Wish you were here: Hindsight Goal Selection for long-horizon dexterous manipulation

Proof Artifact Co-Training for Theorem Proving with Language Models

Language-biased image classification: evaluation based on semantic representations

Practical Integration via Separable Bijective Networks

Robust Learning Meets Generative Models: Can Proxy Distributions Improve Adversarial Robustness?

Autoregressive Quantile Flows for Predictive Uncertainty Estimation

Possibility Before Utility: Learning And Using Hierarchical Affordances

Robbing the Fed: Directly Obtaining Private Data in Federated Learning with Modified Models

Approximation and Learning with Deep Convolutional Models: a Kernel Perspective

Increasing the Cost of Model Extraction with Calibrated Proof of Work

TPU-GAN: Learning temporal coherence from dynamic point cloud sequences

Who Is the Strongest Enemy? Towards Optimal and Efficient Evasion Attacks in Deep RL

RelaxLoss: Defending Membership Inference Attacks without Losing Utility

New Insights on Reducing Abrupt Representation Change in Online Continual Learning

Lossless Compression with Probabilistic Circuits

Programmatic Reinforcement Learning without Oracles

Overcoming The Spectral Bias of Neural Value Approximation

Finding an Unsupervised Image Segmenter in each of your Deep Generative Models

The Geometry of Memoryless Stochastic Policy Optimization in Infinite-Horizon POMDPs

Geometry-Consistent Neural Shape Representation with Implicit Displacement Fields

Deep ReLU Networks Preserve Expected Length

(ends 7:30 PM)

7 p.m.

7:30 p.m.

Town Hall:

Town Hall

(ends 8:30 PM)

9 p.m.

FRI 29 APR

midnight

Invited Talk:

Petascale connectomics and beyond

H. Sebastian Seung

(ends 1:15 AM)

1:30 a.m.

Poster Session 12 [1:30-3:30]

Parallel Training of GRU Networks with a Multi-Grid Solver for Long Sequences

Extending the WILDS Benchmark for Unsupervised Adaptation

On the benefits of maximum likelihood estimation for Regression and Forecasting

Hybrid Random Features

Machine Learning For Elliptic PDEs: Fast Rate Generalization Bound, Neural Scaling Law and Minimax Optimality

Surrogate Gap Minimization Improves Sharpness-Aware Training

Sparse DETR: Efficient End-to-End Object Detection with Learnable Sparsity

Neural Parameter Allocation Search

Discriminative Similarity for Data Clustering

Einops: Clear and Reliable Tensor Manipulations with Einstein-like Notation

SketchODE: Learning neural sketch representation in continuous time

LFPT5: A Unified Framework for Lifelong Few-shot Language Learning Based on Prompt Tuning of T5

Equivariant and Stable Positional Encoding for More Powerful Graph Neural Networks

Model-augmented Prioritized Experience Replay

BadPre: Task-agnostic Backdoor Attacks to Pre-trained NLP Foundation Models

Few-Shot Backdoor Attacks on Visual Object Tracking

NASViT: Neural Architecture Search for Efficient Vision Transformers with Gradient Conflict aware Supernet Training

Transferable Adversarial Attack based on Integrated Gradients

Understanding Domain Randomization for Sim-to-real Transfer

Reliable Adversarial Distillation with Unreliable Teachers

Denoising Likelihood Score Matching for Conditional Score-based Data Generation

Generalized Demographic Parity for Group Fairness

HTLM: Hyper-Text Pre-Training and Prompting of Language Models

Lipschitz-constrained Unsupervised Skill Discovery

Provable Learning-based Algorithm For Sparse Recovery

R5: Rule Discovery with Reinforced and Recurrent Relational Reasoning

Enhancing Cross-lingual Transfer by Manifold Mixup

Model-Based Offline Meta-Reinforcement Learning with Regularization

Huber Additive Models for Non-stationary Time Series Analysis

Reversible Instance Normalization for Accurate Time-Series Forecasting against Distribution Shift

MT3: Multi-Task Multitrack Music Transcription

Relational Surrogate Loss Learning

Neural Spectral Marked Point Processes

A General Analysis of Example-Selection for Stochastic Gradient Descent

Topological Experience Replay

Learning Scenario Representation for Solving Two-stage Stochastic Integer Programs

Online Target Q-learning with Reverse Experience Replay: Efficiently finding the Optimal Policy for Linear MDPs

FALCON: Fast Visual Concept Learning by Integrating Images, Linguistic descriptions, and Conceptual Relations

OntoProtein: Protein Pretraining With Gene Ontology Embedding

KL Guided Domain Adaptation

Acceleration of Federated Learning with Alleviated Forgetting in Local Training

Ada-NETS: Face Clustering via Adaptive Neighbour Discovery in the Structure Space

Reverse Engineering of Imperceptible Adversarial Image Perturbations

Interacting Contour Stochastic Gradient Langevin Dynamics

Self-Joint Supervised Learning

Variational Inference for Discriminative Learning with Generative Modeling of Feature Incompletion

Few-shot Learning via Dirichlet Tessellation Ensemble

Natural Language Descriptions of Deep Features

Shuffle Private Stochastic Convex Optimization

Reinforcement Learning under a Multi-agent Predictive State Representation Model: Method and Theory

TAMP-S2GCNets: Coupling Time-Aware Multipersistence Knowledge Representation with Spatio-Supra Graph Convolutional Networks for Time-Series Forecasting

Out-of-distribution Generalization in the Presence of Nuisance-Induced Spurious Correlations

Universalizing Weak Supervision

On the Generalization of Models Trained with SGD: Information-Theoretic Bounds and Implications

Vector-quantized Image Modeling with Improved VQGAN

Adversarial Robustness Through the Lens of Causality

Sample Efficient Stochastic Policy Extragradient Algorithm for Zero-Sum Markov Game

Do deep networks transfer invariances across classes?

Automatic Loss Function Search for Predict-Then-Optimize Problems with Strong Ranking Property

Constructing Orthogonal Convolutions in an Explicit Manner

QUERY EFFICIENT DECISION BASED SPARSE ATTACKS AGAINST BLACK-BOX DEEP LEARNING MODELS

CrossBeam: Learning to Search in Bottom-Up Program Synthesis

Rethinking Class-Prior Estimation for Positive-Unlabeled Learning

Improving the Accuracy of Learning Example Weights for Imbalance Classification

InfinityGAN: Towards Infinite-Pixel Image Synthesis

Gradient Matching for Domain Generalization

Analyzing and Improving the Optimization Landscape of Noise-Contrastive Estimation

Structure-Aware Transformer Policy for Inhomogeneous Multi-Task Reinforcement Learning

Model Zoo: A Growing Brain That Learns Continually

LoRA: Low-Rank Adaptation of Large Language Models

(ends 3:30 AM)

7 a.m.

Workshop:

Workshop on the Elements of Reasoning: Objects, Structure and Causality

(ends 6:50 PM)

Remarks:

Closing Remarks

(ends 7:30 AM)

8:45 a.m.

Workshop:

Workshop on Agent Learning in Open-Endedness

(ends 8:15 PM)

9 a.m.

Workshop:

AfricaNLP 2022: NLP for African languages

(ends 6:30 PM)

10 a.m.

Workshop:

Wiki-M3L: Wikipedia and Multimodal & Multilingual Research

(ends 8:20 PM)

noon

Workshop:

Gamification and Multiagent Solutions

(ends 10:00 PM)

Workshop:

From Cells to Societies: Collective Learning Across Scales

(ends 10:05 PM)

Workshop:

Geometrical and Topological Representation Learning

(ends 9:00 PM)

Workshop:

Setting up ML Evaluation Standards to Accelerate Progress

(ends 10:35 PM)

Workshop:

Deep Learning for Code

(ends 9:15 PM)

Workshop:

Emergent Communication: New Frontiers

(ends 9:00 PM)

12:45 p.m.

Workshop:

Socially Responsible Machine Learning

(ends 10:00 PM)

Workshop:

GroundedML: Anchoring Machine Learning in Classical Algorithmic Theory

(ends 9:30 PM)

1 p.m.

Workshop:

Deep Generative Models for Highly Structured Data

(ends 7:50 PM)

Workshop:

Deep Learning on Graphs for Natural Language Processing

(ends 10:00 PM)

Workshop:

3rd Workshop on practical ML for Developing Countries: learning under limited/low resource scenarios

(ends 7:45 PM)

Workshop:

Machine Learning for Drug Discovery (MLDD)

(ends 9:30 PM)

3 p.m.

Workshop:

AI for Earth and Space Science

(ends 11:35 PM)

Workshop:

Generalizable Policy Learning in the Physical World

(ends 1:30 AM)

Workshop:

CoSubmitting Summer (CSS) Workshop

(ends 9:00 PM)

4 p.m.

Workshop:

PAIR^2Struct: Privacy, Accountability, Interpretability, Robustness, Reasoning on Structured Data

(ends 1:00 AM)