Byzantine Distributed Source Coding

2025-03-03

AI Summary

This paper investigates robust function reconstruction at the decoder in distributed source coding under Byzantine adversarial attacks, where up to $ s $ users may behave maliciously. We propose the first formal framework for Byzantine-robust distributed source coding that jointly addresses adversarial behavior detection and low-distortion fault-tolerant recovery—overcoming classical modeling assumptions of either fully honest or completely failed nodes. Our approach integrates information-theoretic analysis, random coding construction, consistency checking, and distortion-aware joint decoding to fully characterize the class of reconstructible functions and the achievable information rate region. Theoretically, we prove that in the honest setting, lossless reconstruction is achieved with probability exponentially approaching one; under adversarial conditions, attackers are identified with high probability, or the reconstruction distortion decays exponentially with blocklength.

Technology Category

• Machine Learning: Adversarial Learning & Robustness
• Computer Vision: Adversarial Attacks & Robustness
• Search and Optimization: Distributed Search

Application Category

• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Security and Privacy: Large-scale security measurements

Abstract

We study the distributed source coding problem with $k$ users of which at most $s$ may be controlled by an adversary and characterize the set of functions of the sources the decoder can reconstruct robustly in the following sense - if the users behave honestly, the function is recovered with high probability; if they behave adversarially, with high probability, either one of the adversarial users will be identified or the function is recovered with vanishingly small distortions.

Problem

• Robust distributed source coding with adversarial users
• Characterizing decodable functions under adversarial conditions
• Identifying adversarial users or minimizing function distortions

Innovation

• Byzantine fault-tolerant distributed source coding
• Robust function recovery with adversarial detection
• High probability recovery with minimal distortion