🤖 AI Summary
This paper investigates the efficiency gap between randomized communication protocols and deterministic query algorithms with an Equality oracle in communication complexity. Addressing an open problem posed by Hatami et al., the authors construct an $n$-bit communication problem demonstrating a strong separation: there exists a randomized protocol achieving constant communication cost, whereas any deterministic query algorithm using an Equality oracle requires $n^{Omega(1)}$ queries—yielding the first constant-vs.-polynomial separation of its kind. This result quantifies the inherent difficulty of derandomization in this model and improves upon—and significantly simplifies—the core proof in STOC 2025 concerning the irreducibility of the constant-communication vs. $k$-Hamming-distance hierarchy. Technically, the work integrates tools from randomized communication analysis, query complexity modeling, and oracle-based lower-bound construction, providing new evidence for the fundamental distinction between randomness and determinism in computational models.
📝 Abstract
We exhibit an $n$-bit communication problem with a constant-cost randomized protocol but which requires $n^{Ω(1)}$ deterministic (or even non-deterministic) queries to an Equality oracle. Therefore, even constant-cost randomized protocols cannot be efficiently "derandomized" using Equality oracles. This improves on several recent results and answers a question from the survey of Hatami and Hatami (SIGACT News 2024). It also gives a significantly simpler and quantitatively superior proof of the main result of Fang, Göös, Harms, and Hatami ( STOC 2025), that constant-cost communication does not reduce to the $k$-Hamming Distance hierarchy.