Optimal entanglement-assisted almost-random access codes

Abstract

Entanglement-assisted random access codes (EARACs) compress a large amount of bits, such that any chosen one of them can be recovered with a large probability of success. They have been used for information theoretic treatments of semi-device-independent randomness certification and other communication complexity problems. For high-dimensional problems, the access request distribution can be used to obtain better performance, which until now could only be achieved through numerical methods. This paper makes an analytical account of access request distribution for arbitrary input sizes. We introduce entanglement-assisted almost-random access codes (EAARACs), which maximize the average probability of access under an arbitrary distribution of requests, thus refining EARACs and generalizing the methods used in the study of low-dimensional communication complexity problems. We analytically find the optimal probability of success for EAARACs under arbitrary request distributions and provide an illustrative numerical example.