English

Collaborative Decoding of Polynomial Codes for Distributed Computation

Information Theory 2019-06-03 v1 Distributed, Parallel, and Cluster Computing math.IT

Abstract

We show that polynomial codes (and some related codes) used for distributed matrix multiplication are interleaved Reed-Solomon codes and, hence, can be collaboratively decoded. We consider a fault tolerant setup where tt worker nodes return erroneous values. For an additive random Gaussian error model, we show that for all t<NKt < N-K, errors can be corrected with probability 1. Further, numerical results show that in the presence of additive errors, when LL Reed-Solomon codes are collaboratively decoded, the numerical stability in recovering the error locator polynomial improves with increasing LL.

Keywords

Cite

@article{arxiv.1905.13685,
  title  = {Collaborative Decoding of Polynomial Codes for Distributed Computation},
  author = {Adarsh M. Subramaniam and Anoosheh Heiderzadeh and Krishna R. Narayanan},
  journal= {arXiv preprint arXiv:1905.13685},
  year   = {2019}
}