On Multi-Point Local Decoding of Reed-Muller Codes

Reed-Muller codes are among the most important classes of locally correctable codes. Currently local decoding of Reed-Muller codes is based on decoding on lines or quadratic curves to recover one single coordinate. To recover multiple coordinates simultaneously, the naive way is to repeat the local decoding for recovery of a single coordinate. This decoding algorithm might be more expensive, i.e., require higher query complexity. In this paper, we focus on Reed-Muller codes with evaluation polynomials of total degree d . σ √ q for some σ ∈ (0, 1). By introducing a local decoding of Reed-Muller codes via the concept of codex that has been used for arithmetic secret sharing [6, 7], we are able to locally recover arbitrarily large number k of coordinates simultaneously at the cost of querying O(k √ q) coordinates, where q is the code alphabet size. It turns out that our local decoding of ReedMuller codes shows (perhaps surprisingly) that accessing k locations is in fact cheaper than repeating the procedure for accessing a single location for k times. In contrast, by repetition of local decoding for recovery of a single coordinate, one has to query Ω(k √ q log k/ log q) coordinates for k = q √ q) (and query O(kq) coordinates for k = q √ , respectively). Furthermore, our decoding success probability is 1− with = O (( 1 √ q )k) . To get the same success probability from repetition of local decoding for recovery of a single coordinate, one has to query O(k √ q log k/ log q) coordinates (or O(kq) coordinates for k = q √ , respectively). In addition, our local decoding also works for recovery of one single coordinate as well and it gives a better success probability than the one by repetition of local decoding on curves. The main tool to realize codex is based on algebraic function fields (or more precisely, algebraic geometry codes). Our estimation of success error probability is based on error probability bound for t-wise linearly independent variables given in [2]. ∗CWI, Amsterdam and Mathematical Institute, Leiden University (email: [email protected]) †School of Physical and Mathematical Sciences, Nanyang Technological University, Singapore (email: [email protected]) ‡School of Physical and Mathematical Sciences, Nanyang Technological University, Singapore (email: [email protected]) 1 ISSN 1433-8092 Electronic Colloquium on Computational Complexity, Report No. 51 (2016)