The LAMA-WeST Lab (Web, Semantics and Text) conducts fundamental research in knowledge-centric artificial intelligence, focusing on how language models can represent and reason over structured knowledge. Its work explores the integration of neural and symbolic AI, with a central emphasis on mapping natural language to formal representations (e.g., SPARQL) for principled reasoning over knowledge graphs and ontologies.
A key direction is the study of nonparametric large language models (LLMs), where knowledge is dynamically retrieved from external sources rather than stored solely in model parameters. This enables more interpretable, controllable, and grounded generation, while raising core questions about memory, inference, and generalization.
The lab also investigates the foundations of LLM safety and reliability, including knowledge verification and factuality checking. Its research addresses hallucination reduction, bias mitigation, and the validation of outputs against structured knowledge, contributing to more transparent, robust, and trustworthy AI systems.
Application domains include law, medicine, and oncology, which serve as testbeds for developing precise, explainable, and verifiable interactions between language and complex knowledge systems.
Our lab is funded through government and industry institutions including NSERC, FRQS, SSHRC, IVADO and MITACS. Our students are also part of MILA.
Led by Prof. Amal Zouaq, the LAMA-WeST lab is welcoming new collaborations in the academic and industrial community.
News
Aditya Sharma is presenting his paper “Reducing hallucinations in language model-based SPARQL query generation using post-generation memory retrieval” at EACL 2026 (Findings).
We have just launched our new website! Explore the new design and updated content across all sections.
Congratulations to Alexis Brissard for successfully defending his thesis!
Featured Seminars
Our Research
Our research explores natural language processing, knowledge graphs and machine learning, with applications in education, healthcare and intelligent information systems.
Our Seminars
We host regular seminars where lab members present ongoing projects, new results and other topics in natural language processing and AI.
Our Team
Our team brings together students, postdoctoral researchers and collaborators interested in natural language processing, semantic web technologies and applied AI.