Observation and Distinction. Representing Information in Infinite Games
Dietmar Berwanger, Laurent Doyen

TL;DR
This paper compares two finite-state automata-based models for representing imperfect information in infinite games, showing the greater expressiveness of indistinguishability relations and providing a decidable characterization and construction method.
Contribution
It introduces a formal comparison between observation-based and indistinguishability-based models, establishing the latter's superior expressiveness and providing a decision procedure for finite-state representation.
Findings
Indistinguishability relations are strictly more expressive than observation-based models.
A decidable characterization of representable indistinguishability relations is provided.
A procedure to construct a finite-state observation function from indistinguishability relations is developed.
Abstract
We compare two approaches for modelling imperfect information in infinite games by using finite-state automata. The first, more standard approach views information as the result of an observation process driven by a sequential Mealy machine. In contrast, the second approach features indistinguishability relations described by synchronous two-tape automata. The indistinguishability-relation model turns out to be strictly more expressive than the one based on observations. We present a characterisation of the indistinguishability relations that admit a representation as a finite-state observation function. We show that the characterisation is decidable, and give a procedure to construct a corresponding Mealy machine whenever one exists.
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsComputability, Logic, AI Algorithms · Logic, programming, and type systems · Game Theory and Applications
