Communication over an Arbitrarily Varying Channel under a State-Myopic Encoder
Amitalok J. Budkuley, Sidharth Jaggi

TL;DR
This paper investigates the capacity of a communication system over an arbitrarily varying channel with a state known non-causally to the encoder through a noisy observation, providing capacity characterizations under different randomization scenarios.
Contribution
It characterizes the capacity of a state-dependent AVC with a noisy state observation at the encoder, including cases with shared randomness and encoder-only randomization.
Findings
Capacity with shared randomness is characterized.
Capacity remains unchanged with encoder-only randomization when positive.
Special cases of the model are also analyzed.
Abstract
We study the problem of communication over a discrete arbitrarily varying channel (AVC) when a noisy version of the state is known non-causally at the encoder. The state is chosen by an adversary which knows the coding scheme. A state-myopic encoder observes this state non-causally, though imperfectly, through a noisy discrete memoryless channel (DMC). We first characterize the capacity of this state-dependent channel when the encoder-decoder share randomness unknown to the adversary, i.e., the randomized coding capacity. Next, we show that when only the encoder is allowed to randomize, the capacity remains unchanged when positive. Interesting and well-known special cases of the state-myopic encoder model are also presented.
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
