TY - GEN
T1 - Locally Decodable Slepian-Wolf Compression
AU - Vatedka, Shashank
AU - Chandar, Venkat
AU - Tchamkerten, Aslan
N1 - Publisher Copyright:
© 2022 IEEE.
PY - 2022/1/1
Y1 - 2022/1/1
N2 - This paper investigates the Slepian-Wolf distributed compression of two sources Xn and Yn with the additional property that any pair (Xi, Yi) should reliably be decoded by probing a small number d of compressed bits. We show that for certain source distributions, the error probability of any such local decoder is lower bounded by 2-O(d), in the worst case over index i, whenever one of the sources is compressed below its entropy. Unlike the single-source setup, it is thus impossible to simultaneously achieve constant local decodability d and vanishing local decoding error probability as n increases. We also provide a compression scheme with a local decoder that almost achieves the above lower bound.
AB - This paper investigates the Slepian-Wolf distributed compression of two sources Xn and Yn with the additional property that any pair (Xi, Yi) should reliably be decoded by probing a small number d of compressed bits. We show that for certain source distributions, the error probability of any such local decoder is lower bounded by 2-O(d), in the worst case over index i, whenever one of the sources is compressed below its entropy. Unlike the single-source setup, it is thus impossible to simultaneously achieve constant local decodability d and vanishing local decoding error probability as n increases. We also provide a compression scheme with a local decoder that almost achieves the above lower bound.
U2 - 10.1109/ISIT50566.2022.9834371
DO - 10.1109/ISIT50566.2022.9834371
M3 - Conference contribution
AN - SCOPUS:85129486430
T3 - IEEE International Symposium on Information Theory - Proceedings
SP - 1430
EP - 1435
BT - 2022 IEEE International Symposium on Information Theory, ISIT 2022
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 2022 IEEE International Symposium on Information Theory, ISIT 2022
Y2 - 26 June 2022 through 1 July 2022
ER -