Search

We study the problem of one-shot channel simulation of DMCs with unlimited shared randomness. For any fixed tolerance measured in total variational distance, we propose an achievability bound and a converse bound on the size of the code to simulate the channel. The achievability bound utilizes the convex split lemma, whereas the converse bound is the result of the relationships between smoothed max-divergences and the max-mutual information. The achievability proof does not rely on a "universal state" (compared with some previous related works), and provides a tighter bound. Using the two bounds, we also provide an alternative proof to the reverse Shannon theorem.

We study channel simulation under common randomness-assistance in the finite-blocklength regime and identify the smooth channel max-information as a linear program one-shot converse on the minimal simulation cost for fixed error tolerance. We show that this one-shot converse can be achieved exactly using no-signaling assisted codes, and approximately achieved using common randomness-assisted codes. Our one-shot converse thus takes on an analogous role to the celebrated meta-converse in the complementary problem of channel coding, and find tight relations between these two bounds. We asymptotically expand our bounds on the simulation cost for discrete memoryless channels, leading to the second-order as well as the moderate deviation rate expansion, which can be expressed in terms of the channel capacity and channel dispersion known from noisy channel coding. Our techniques extend to discrete memoryless broadcast channels. In stark contrast to the elusive broadcast channel capacity problem, we show that the reverse problem of broadcast channel simulation under common randomness-assistance allows for an efficiently computable single-letter characterization of the asymptotic rate region in terms of the broadcast channel's multi-partite mutual information. Finally, we present a Blahut-Arimoto type algorithm to compute the rate region efficiently., Comment: 32 pages, 10 figures

Quantum state transfer involves two parties who use pre-shared entanglement and noiseless communication in order to transfer parts of a quantum state. In this work, we quantity the communication cost of one-shot state splitting in terms of the partially smoothed max-information. We then give an analysis of state splitting in the moderate deviation regime, where the error in the protocol goes sub-exponentially fast to zero as a function of the number of i.i.d. copies. The main technical tool we derive is a tight relation between the partially smoothed max-information and the hypothesis testing relative entropy, which allows us to obtain the expansion of the partially smoothed max-information for i.i.d. states in the moderate deviation regime. This then also establishes the moderate deviation analysis for other variants of state transfer such as state merging and source coding.

The moderate deviation regime is concerned with the finite block length trade-off between communication cost and error for information processing tasks in the asymptotic regime, where the communication cost approaches a capacity-like quantity and the error vanishes at the same time. We find exact characterisations of these trade-offs for a variety of fully quantum communication tasks, including quantum source coding, quantum state splitting, entanglement-assisted quantum channel coding, and entanglement-assisted quantum channel simulation. The main technical tool we derive is a tight relation between the partially smoothed max-information and the hypothesis testing relative entropy. This allows us to obtain the expansion of the partially smoothed max-information for i.i.d. states in the moderate deviation regime., Comment: 32 pages