📄📄📄 3 Papers accepted @ FLoC 2026!! 🎉🎉🎉

Sat, 30 May 2026 00:00:00 +0000

I’m very excited to share that three of our papers have been accepted at FLoC 2026! 🎉

Two papers were accepted to the 42nd International Conference on Logic Programming ( ), while a third paper was accepted to the LLMs meet Constraint Solving ( ) Workshop.

📚 Can Automated Feedback Turn Students into Happy Prologians?

Providing personalised feedback is crucial for effective learning, but delivering it at scale remains challenging. In this paper, we present ProHelp, an automated assessment platform for Prolog built on top of , and evaluate it through a large-scale deployment in an undergraduate logic programming course. Our results show that students perceive automated feedback as highly valuable, with automatic testing, open-choice-point warnings, and predicate-level scoring emerging as the most useful feedback mechanisms.

Read the ProHelp paper

Ricardo Brancas, Pedro Orvalho, Carolina Carreira, Vasco Manquinho, Ruben Martins (2026). Can Automated Feedback Turn Students into Happy Prologians?. In ICLP 2026.
PDF project Preprint DOI

🐞 What Bugs Do Prolog Students Write? An Empirical Taxonomy and Data-Driven Mutation Framework

Automated feedback and repair systems require realistic bug datasets that reflect the mistakes students actually make. In this paper, we analyse 7,201 Prolog submissions from 265 students to construct a detailed taxonomy of Prolog programming errors. Based on this empirical study, we introduce LogMorph, a data-driven mutation framework that generates realistic buggy Prolog programs according to the observed distribution of student mistakes, enabling more representative evaluation of debugging, repair, and educational tools.

Read the LogMorph paper

Ricardo Brancas, Pedro Orvalho, Carolina Carreira, Vasco Manquinho, Ruben Martins (2026). What Bugs Do Prolog Students Write? An Empirical Taxonomy and Data-Driven Mutation Framework. In ICLP 2026.
project

🤖 Solving MaxSAT Problems from Natural Language Descriptions with LLMs and PySAT

Can large language models make formal optimisation technologies accessible through natural language? In this workshop paper, we explore a neuro-symbolic approach in which an LLM translates a natural language problem description into executable PySAT code that constructs and solves a Maximum Satisfiability (MaxSAT) instance. By combining the flexibility of LLMs for semantic interpretation with the reliability of exact MaxSAT solvers, the approach significantly improves over direct LLM-based problem solving while maintaining formal correctness guarantees.

Read the paper

Pedro Orvalho, Marta Kwiatkowska, Guillem Alenyà, Felip Manyà (2026). Solving MaxSAT Problems from Natural Language Descriptions with LLMs and PySAT. In LLM-Solve @ FLoC 2026.
PDF project Preprint

I’m very grateful to all my collaborators!

See you in Lisbon @ FLoC 2026!!

Solving MaxSAT Problems from Natural Language Descriptions with LLMs and PySAT

Wed, 27 May 2026 00:00:00 +0000

Neuro-Symbolic AI | Pedro Orvalho

📄📄📄 3 Papers accepted @ FLoC 2026!! 🎉🎉🎉

📚 Can Automated Feedback Turn Students into Happy Prologians?

🐞 What Bugs Do Prolog Students Write? An Empirical Taxonomy and Data-Driven Mutation Framework

🤖 Solving MaxSAT Problems from Natural Language Descriptions with LLMs and PySAT

Solving MaxSAT Problems from Natural Language Descriptions with LLMs and PySAT