Talk Title: Codes for the Input-Constrained Binary Erasure Channel
Speaker: Mr. Arvind Rameshwar
Date and Time: June 25, 2021 (Friday), 5:00 PM to 5:45 PM
Talk Recording YouTube Link: https://youtu.be/JJH2k7ZuCfU
Abstract:
In this talk, we take up the familiar framework of a discrete memoryless channel, commonly used as a model for noisy communication and storage, with constrained inputs. In particular, we consider the binary erasure channel (BEC) with inputs that obey the constraint that there are at least d zeros between every pair of successive ones. Such constrained sequences find application in the mitigation of intersymbol interference and synchronization errors in storage media such as hard disks and flash memories, and arise as natural constraints in DNA storage and computation.
We first study the setting when there is noiseless feedback from the decoder and derive an explicit expression for the feedback capacity. We also demonstrate a simple feedback capacity-achieving coding scheme, that draws inspiration from the simple “repeat-until-success” coding scheme that achieves the feedback capacity of an unconstrained BEC.
We then move to a brief discussion on the capacity without feedback when d=1. We provide the motivation behind a coding scheme using Markov inputs, which arises from a two-timescale stochastic approximation algorithm and which achieves numerical values of rates very close to capacity.
The talk is based on joint work with Aashish Tolambiya and Prof. Navin Kashyap.
Biography:
Arvind Rameshwar received the B.E. (Hons.) degree in Electronics and Communication Engineering from BITS Pilani University, India, in 2018, and was the gold medallist of the graduating class of 2018. He is currently pursuing the Ph.D. degree at the Department of Electrical Communication Engineering, at the Indian Institute of Science, Bengaluru. He is a recipient of the Prime Minister’s Research Fellowship 2020 and was part of a team that won a Qualcomm Innovation Fellowship India 2020. His research interests lie in information theory and coding for finite-state channels.
