Interactive Coding with Small Memory and Improved Rate

TL;DR

This paper presents a method to convert any adaptive interactive protocol into a noise-robust version that uses small memory and maintains near-optimal communication rates, even under adversarial noise.

Contribution

It introduces a transformation that makes protocols resilient to adversarial noise with limited memory, achieving near-optimal communication complexity.

Findings

Protocols can be made robust to adversarial noise with small memory.

The transformed protocol has only a small increase in rounds of communication.

Memory usage is polylogarithmic in the original protocol size.

Abstract

In this work, we study two-party interactive coding for adversarial noise, when both parties have limited memory. We show how to convert any adaptive protocol $\Pi$ into a protocol $\Pi'$ that is robust to an $\epsilon$-fraction of adversarial corruptions, not too much longer than $\Pi$, and which uses small space. More precisely, if $\Pi$ requires space $\log(s)$ and has $|\Pi|$ rounds of communication, then $\Pi'$ requires $O_\epsilon(\log s \log |\Pi|)$ memory, and has $$|\Pi'| = |\Pi|\cdot\left( 1 + O\left( \sqrt{ \epsilon \log \log 1/\epsilon } \right)\right)$$ rounds of communication. The above matches the best known communication rate, even for protocols with no space restrictions.