Context & Goals

Answer Set Programming (ASP) is a well-established logic-based paradigm in Knowledge Representation and Reasoning (KRR), widely applied in both academic and industrial contexts. The success of ASP relies on the combination of a declarative modeling language with the availability of efficient solving technology, which make ASP ideal for tackling hard combinatorial problems with exponentially large search spaces.

With this tutorial we aim at guiding participants through the whole ASP landscape, from the ASP essential building blocks to recent advancements which empower both efficiency and expressiveness. We will also highlight ASP’s evolving role in the broader KRR landscape, including its synergy with modern Large Language Models (LLMs): from one hand ASP represent a valid solution to strengthen the reasoning capabilities of LLMs; while on the other hand, LLMs may offer a more accessible way to generate ASP specification.

To this end, we will start by introducing the ASP language with several examples at a practical level, showing how to use ASP for modeling and solving combinatorial search problems with ASP. In particular, we first introduce the guess-check (a.k.a., generate-test) which represents the classical ASP programming methodology followed to model NP-complete problems. Then, the internal principles of modern ASP solvers are also overviewed, to have an idea of how programs are evaluated and what are the critical aspects of encodings which might affect the performance of the evaluation.

In the second part, we explore recent extensions of ASP proposed for modeling problems beyond NP, with a particular focus on the ASP with Quantifiers (ASP(Q)) formalism, along with intuitive modeling examples.

Finally, we discuss emerging neuro-symbolic approaches that integrate ASP with Large Language Models (LLMs): ASP can enhance the reasoning capabilities of LLMs, while LLMs can assist ASP developers by automatically generating ASP code from textual problem descriptions, demonstrating the growing synergy between symbolic reasoning and modern AI technologies. 

The tutorial aims to be very interactive, participants will be pushed to solve problems with ASP by using ASP Chef.

Tutorial Outline

• Introduction to ASP

• Modeling and solving problems with ASP

• ASP Solving basics

• Solving hard combinatorial optimization problems with ASP

• Combining ASP with Large Language Models

Speakers