BEGIN:VCALENDAR VERSION:2.0 PRODID:researchseminars.org CALSCALE:GREGORIAN X-WR-CALNAME:researchseminars.org BEGIN:VEVENT SUMMARY:Cheuk Ting LI (Chinese University of Hong Kong) DTSTART:20260923T150000Z DTEND:20260923T161500Z DTSTAMP:20261006T143522Z UID:AAIT/54 DESCRIPTION:Title: A Correspondence between Coding Problems and Logic via Confusion Hypergraphs .\nby Cheuk Ting LI (Chinese University of Hong Kong) as part of Semin ar on Algorithmic Aspects of Information Theory\n\n\nAbstract\nInformation is often represented as random variables or partitions of the sample spac e\, where we can take the "conjunction" of two pieces of information as th e joint random variable\, but cannot perform other logical operations. By generalizing partitions to confusion hypergraphs (which form a Heyting alg ebra)\, we can take conjunction\, disjunction and implication between info rmation. In this talk\, we will discuss connections between confusion hype rgraphs and logic (generalized inquisitive logic and intuitionistic logic) \, and show how we can directly compute an almost-optimal coding scheme fo r various coding settings (e.g.\, network coding and index coding) by tran slating the settings into logical formulae\, and evaluating the formulae o ver hypergraphs. The optimal communication cost is given approximately by the hypergraph entropy. We will also discuss several identities and inequa lities on hypergraph entropy\, and their interpretations in coding setting s.\n LOCATION:https://researchseminars.org/talk/AAIT/54/ END:VEVENT END:VCALENDAR