On relating one-way classical and quantum communication complexities

Abstract

Communication complexity is the amount of communication needed to compute a function when the function inputs are distributed over multiple parties. In its simplest form, one-way communication complexity, Alice and Bob compute a function $f(x,y)$, where $x$ is given to Alice and $y$ is given to Bob, and only one message from Alice to Bob is allowed. A fundamental question in quantum information is the relationship between one-way quantum and classical communication complexities, i.e., how much shorter the message can be if Alice is sending a quantum state instead of bit strings? We make some progress towards this question with the following results. Let $f: \mathcal{X} \times \mathcal{Y} \rightarrow \mathcal{Z} \cup \{\bot\}$ be a partial function and $μ$ be a distribution with support contained in $f^{-1}(\mathcal{Z})$. Denote $d=|\mathcal{Z}|$. Let $\mathsf{R}^{1,μ}_ε(f)$ be the classical one-way communication complexity of $f$; $\mathsf{Q}^{1,μ}_ε(f)$ be the quantum one-way communication complexity of $f$ and $\mathsf{Q}^{1,μ, *}_ε(f)$ be the entanglement-assisted quantum one-way communication complexity of $f$, each with distributional error (average error over $μ$) at most $ε$. We show: 1) If $μ$ is a product distribution, $η> 0$ and $0 \leq ε\leq 1-1/d$, then, $$\mathsf{R}^{1,μ}_{2ε-dε^2/(d-1)+ η}(f) \leq 2\mathsf{Q}^{1,μ, *}_ε(f) + O(\log\log (1/η))\enspace.$$ 2)If $μ$ is a non-product distribution and $\mathcal{Z}=\{ 0,1\}$, then $\forall ε, η> 0$ such that $ε/η+ η< 0.5$, $$\mathsf{R}^{1,μ}_{3η}(f) = O(\mathsf{Q}^{1,μ}_{ε}(f) \cdot \mathsf{CS}(f)/η^3)\enspace,$$ where \[\mathsf{CS}(f) = \max_{y} \min_{z\in\{0,1\}} \vert \{x~|~f(x,y)=z\} \vert \enspace.\]