ISIT 2024

206 papers
The 2024 IEEE International Symposium on Information Theory (ISIT) featured 192 papers addressing core information-theoretic concepts and their applications in machine learning and quantum computing. Research themes included the analysis of message passing via density evolution, spatial coupling for closing the BP-MAP gap, and the development of sparse graph codes for group testing. Attention was directed toward the intersection of information theory and statistical learning, with work focusing on language model alignment, inference for unnormalized statistical models, and regularizing empirical risk using f-divergences. The 2024 edition also highlighted emerging topics such as algorithmic fairness and quantum information. Researchers explored methods for auditing intersectional subgroup fairness and decomposing algorithmic discrimination for targeted interventions. Quantum-focused contributions addressed entanglement costs in state discrimination and verification through quantum rigidity. Prolific authors contributing to these developments included Zhongyi Huang, Bo Bai, Deniz Gunduz, and Shlomo Shamai. The proceedings further included advancements in private vector database querying and high-dimensional linear regression models.

Top Research Directions

Research directions most represented by ISIT 2024 papers.

Characterizing Capacity Achieving Input Distributions under Physical Constraints 2
Optimizing Repair Bandwidth for Distributed Erasure Codes 2
Learning Minimal Sufficient Representations via Information Bottlenecks 2
Information Theoretic Fair Representation Learning 2
Tightening Theoretical Bounds by Optimizing Constant Factors 2
The Empirical Validation Gap 2
Group Testing with Sum and Quantized Observations 1
Non-Monotonic Information Aging in Remote Inference 1
Rate Distortion of Asynchronously Sampled Cyclostationary Processes 1
Universal Source Coding for Unknown Distributions 1
Bypassing Exponential Regret with Polynomial Distance Penalties 1
Linear Regression without Correspondence 1
Analyzing Information Limits via the Information Spectrum 1
Slack Matrix Factorization and Polyhedral Complexity 1
Proving Statistical Consistency for Complex Estimators 1

Top Authors

AuthorPapers
Alex Dytso5
Deniz Gunduz5
Ashish J Khisti4
Bo Bai4
Cheuk Ting Li4
Rawad Bitar4
Shahab Asoodeh4
Antonia Wachter-Zeh3
Dor Tsur3
Flavio Calmon3
Gong Zhang3
H. Vincent Poor3
Jie Wang3
Linqi Song3
Oliver Kosut3
Vincent Y. F. Tan3
Abdellatif Zaidi2
Alexandre Graell i Amat2
Ali Tajer2
Anuran Makur2

Impactful Papers

Papers from ISIT 2024 that shaped research clusters.

Influential

Papers

Title
Prediction with Noisy Expert Advice
Private Approximate Nearest Neighbor Search for Vector Database Querying
Proving Information Inequalities by Gaussian Elimination
Quantum Soft Covering and Decoupling with Relative Entropy Criterion
Quickest Detection in High-Dimensional Linear Regression Models via Implicit Regularization
Rate-Distortion-Perception Tradeoff for Lossy Compression Using Conditional Perception Measure
Rate-Limited Optimal Transport for Quantum Gaussian Observables
Reed-Solomon Codes over Cyclic Polynomial Ring with Lower Encoding/Decoding Complexity
Regularized Linear Regression for Binary Classification
Rejection-Sampled Universal Quantization for Smaller Quantization Errors
Repairing a Single Erasure in Reed-Solomon Codes with Side Information
Repairing Schemes for Tamo-Barg Codes
Repairing with Zero Skip Cost
Robust Distributed Gradient Descent to Corruption over Noisy Channels
Robust Multi-Hypothesis Testing with Moment-Constrained Uncertainty Sets
Robust Semi-supervised Learning via f-Divergence and α- Renyi Divergence
Robust Semi-supervised Learning via f-Divergence and α-Rényi Divergence
Second-Order Characterization of Minimax Parameter Estimation in Restricted Parameter Space
Section-Wise Revolving NBP-Like Decoders for QC-LDPC Codes
Secure Inference for Vertically Partitioned Data Using Multiparty Homomorphic Encryption
Secure Storage Using Maximally Recoverable Locally Repairable Codes
Secure Submodel Aggregation for Resource-Aware Federated Learning
Set Transformation: Trade-Off Between Repair Bandwidth and Sub-Packetization
Several Interpretations of Max-Sliced Mutual Information
Shared Information Under Simple Markov Independencies
Sharp Analysis of Out-of-Distribution Error for "Importance-Weighted" Estimators in the Overparameterized Regime
Sharper Rates of Convergence for the Tensor Graphical Lasso Estimator
Sharp Information-Theoretic Thresholds for Shuffled Linear Regression
SignSGD-FV: Communication-Efficient Distributed Learning Through Heterogeneous Edges
Simultaneous Computation and Communication over MAC
Sliding Secure Symmetric Multilevel Diversity Coding
Sparse Gaussian Gradient Code
Sparsity-Based Secure Gradient Aggregation for Resource-Constrained Federated Learning
Spectral Convergence of Simplicial Complex Signals
Statistic Maximal Leakage
Subgraph Matching via Partial Optimal Transport
Subset Adaptive Relaying for Streaming Erasure Codes
Supervised Contrastive Representation Learning: Landscape Analysis with Unconstrained Features
TexShape: Information Theoretic Sentence Embedding for Language Models
The Bethe Partition Function and the SPA for Factor Graphs based on Homogeneous Real Stable Polynomials
The Error Probability of Spatially Coupled Sparse Regression Codes over Memoryless Channels
Theoretical Guarantees of Data Augmented Last Layer Retraining Methods
The Rate-Distortion-Perception Trade-off: the Role of Private Randomness
Thermal-Aware Channel with Multiple Wires
Threshold Saturation for Quantitative Group Testing with Low-Density Parity-Check Codes
Throughput and Latency of Network Coding in Line Networks with Outages
Tight Lower Bound on Cross-Rack Update Bandwidth and Explicit Constructions
Towards General Function Approximation in Nonstationary Reinforcement Learning
Towards Optimal Inverse Temperature in the Gibbs Algorithm
Unequal Message Protection: One-Shot analysis via Poisson Matching Lemma

« prev Page 4/5 next »