Solving Epistemic Logic Programs Using Generate-and-Test with Propagation

Authors: Jorge Fandinno, Lute Lillo

AAAI 2025 | Venue PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We implement a new solver based on these theoretical findings and experimentally show that it outperforms existing solvers by achieving a 3.3x speed-up and solving 91% more instances on well-known benchmarks.
Researcher Affiliation Academia Jorge Fandinno1, Lute Lillo1,2 1University of Nebraska Omaha, Omaha, NE, USA 2University of Vermont, Burlington, VT, USA EMAIL EMAIL
Pseudocode Yes Algorithm 1: Generate-and-test computation of n worldviews of a program Π in normal form.
Open Source Code Yes This is released as a new version of eclingo (https://github.com/potassco/eclingo).
Open Datasets Yes For the experimental evaluation, we use the well-established benchmark suite by Son et al. (2017). It consists of three problems: the Eligibility problem that represents reasoning in disjunctive databases (Gelfond 1991), and the Yale Shooting and Bomb in the Toilet problems that are instances of conformant planning.
Dataset Splits No The paper mentions expanding a benchmark suite, for example, by increasing the number of students or packages, but does not specify how these datasets are split into training, validation, or test sets for reproduction purposes.
Hardware Specification Yes We set a timeout of 600s and ran our experiments on a machine powered by an Intel Core Processor (Broadwell) with 12 CPUs running at 2095.078 MHz and 100GB of RAM. The OS is Red Hat Enterprise Linux Server 7.9.
Software Dependencies Yes Our implementation is built on top of version 5.7 of the ASP solver clingo (Gebser et al. 2019) using Python and ASP.
Experiment Setup Yes We set a timeout of 600s and ran our experiments on a machine powered by an Intel Core Processor (Broadwell) with 12 CPUs running at 2095.078 MHz and 100GB of RAM.