CAIML doctoral researcher Alexander Beiser presented “FastFound” at KR 2025 in Melbourne, introducing a new predicate-decoupled grounding technique that significantly mitigates the longstanding grounding bottleneck in Answer Set Programming.

Abstract

At KR 2025, the 22nd International Conference on Principles of Knowledge Representation and Reasoning held in Melbourne, Australia (11–17 November 2025), CAIML doctoral researcher Alexander Beiser presented the main-track paper “FastFound: Easing the ASP Bottleneck via Predicate-Decoupled Grounding” in the ASP session.

FastFound is a joint work by Alexander Beiser (TU Wien), Martin Gebser (University of Klagenfurt), Markus Hecher (CNRS / CRIL, Université d’Artois), and Stefan Woltran (TU Wien). The paper tackles the long-standing grounding bottleneck in Answer Set Programming (ASP), which arises from the combinatorial blow-up during variable instantiation in traditional ground-and-solve systems.

Building on the Body-Decoupled Grounding (BDG) paradigm, the authors introduce FastFound, an alternative foundedness check that grounds each predicate separately and enforces foundedness implicitly. This yields a quadratic reduction in grounding size compared to previous BDG encodings and achieves an essentially optimal bound for tight normal programs.

Proceedings

Proceedings link: FastFound: Easing the ASP Bottleneck via Predicate-Decoupled Grounding