Coding for Interactive Communication" is the study of making interactive protocols resilient to errors. But we can ask the following question: what if the original protocol was only sending a single message from Alice to Bob ? Can using interaction (allowing Bob to send noisy feedback) help them achieve higher error-resilience than what is possible with error correcting codes ?
The talk will be based on the paper "Interactive Error Correcting Codes Over Binary Erasure Channels Resilient to > 1/2 Adversarial Corruption" by Meghal Gupta, Yael Tauman Kalai and Rachel Yun Zhang.
Reference: https://arxiv.org/abs/2111.04181.
