Ítem


Combining Constraint Programming Reasoning with Large Language Model Predictions

Constraint Programming (CP) and Machine Learning (ML) face challenges in text generation due to CP’s struggle with implementing "meaning" and ML’s difficulty with structural constraints. This paper proposes a solution by combining both approaches and embedding a Large Language Model (LLM) in CP. The LLM handles word generation and meaning, while CP manages structural constraints. This approach builds on On-the-fly Constraint Programming Search (OTFS), improving it using LLM-generated domains. Compared to Beam Search (BS), a standard NLP method, this combined approach (OTFS with LLM) is faster and produces better results, ensuring all constraints are satisfied. This fusion of CP and ML presents new possibilities for enhancing text generation under constraints

7763.mp4 7763.mp3

Universitat de Girona. Departament d’Informàtica, Matemàtica Aplicada i Estadística

Altres contribucions: Universitat de Girona. Departament d’Informàtica, Matemàtica Aplicada i Estadística
Autor: Régin, Florian
De Maria, Elisabetta
Bonlaron, Alexandre
Data: 3 setembre 2024
Resum: Constraint Programming (CP) and Machine Learning (ML) face challenges in text generation due to CP’s struggle with implementing "meaning" and ML’s difficulty with structural constraints. This paper proposes a solution by combining both approaches and embedding a Large Language Model (LLM) in CP. The LLM handles word generation and meaning, while CP manages structural constraints. This approach builds on On-the-fly Constraint Programming Search (OTFS), improving it using LLM-generated domains. Compared to Beam Search (BS), a standard NLP method, this combined approach (OTFS with LLM) is faster and produces better results, ensuring all constraints are satisfied. This fusion of CP and ML presents new possibilities for enhancing text generation under constraints
7763.mp4 7763.mp3
Format: audio/mpeg
video/mp4
Accés al document: http://hdl.handle.net/10256.1/7763
Llenguatge: eng
Editor: Universitat de Girona. Departament d’Informàtica, Matemàtica Aplicada i Estadística
Col·lecció: 30th International Conference on Principles and Practice of Constraint Programming
Drets: Attribution-NonCommercial-ShareAlike 4.0 International
URI Drets: http://creativecommons.org/licenses/by-nc-sa/4.0/
Matèria: Programació per restriccions (Informàtica) -- Congressos
Constraint programming (Computer science) -- Congresses
Aprenentatge automàtic -- Congressos
Machine learning -- Congresses
Títol: Combining Constraint Programming Reasoning with Large Language Model Predictions
Tipus: info:eu-repo/semantics/lecture
Repositori: DUGiMedia

Matèries

Autors