Universal graph representation of stabilizer codes
Jonathan Z. Lu,Andrey Boris Khesin,Peter W. Shor
2024-11-08
Abstract:We introduce a representation of $[[n, k]]$ stabilizer codes as semi-bipartite graphs wherein $k$ ``input'' nodes map to $n$ ``output'' nodes, such that output nodes may connect to each other but input nodes may not. We prove that this graph representation is in bijection with tableaus and give an efficient compilation algorithm that transforms tableaus into graphs. We then show that this map is efficiently invertible, which gives a new universal recipe for code construction by way of finding graphs with sufficiently nice properties.
The graph representation gives insight into both code construction and algorithms. To the former, we argue that graphs provide a flexible platform for building codes particularly at smaller (non-asymptotic) scales. We construct as examples constant-size codes, e.g. a $[[64, 6, 5]]$ code and a family of roughly $[[n, \frac{n}{\log n}, \log n]]$ codes. We also leverage graphs in a probabilistic analysis to extend the quantum Gilbert-Varshamov bound into a three-way distance-rate-weight tradeoff. To the latter, we show that key coding algorithms -- distance approximation, weight reduction, and decoding -- are unified as instances of a single optimization game on a graph. Moreover, key code properties such as distance, weight, and encoding circuit depth, are all controlled by the graph degree. We give efficient algorithms for producing simple encoding circuits whose depths scale as twice the degree and for implementing logical diagonal and certain Clifford gates with non-constant but reduced depth. Finally, we construct a simple efficient decoding algorithm and prove a performance guarantee for a certain class of graphs, including the roughly $[[n, \frac{n}{\log n}, \log n]]$ code. These results give evidence that graphs are generically useful for the study of stabilizer codes and their practical implementations.
Quantum Physics