Nicolas Gontier
Research Scientist
AI Frontier Research
Nicolas is a Research Scientist in the Conversational Agent Team (CAT) at ServiceNow Research. He obtained his Ph.D. in 2023 at the Mila institute and Polytechnique Montreal supervised by professor Christopher Pal. His research interest lies at the intersection of Natural Language Processing and multi-step reasoning (or as he likes to call it “Natural Language Reasoning”). Prior to his PhD, he did his Master’s degree at McGill University with Joelle Pineau during which he worked on Dialog Systems.
Intérêts
- Natural Language Processing
- Dialog
- Large Language Models
- Text2Code
- Compositional Generalization
- Multi Step Reasoning
Publications
How to Train Your LLM Web Agent: A Statistical Diagnosis.
Neural Information Processing Systems (NeurIPS),
2025.
How to Train Your LLM Web Agent: A Statistical Diagnosis (Oral).
Workshop at the International Conference of Machine Learning (ICML),
2025.
Seq-VCR: Preventing Collapse in Intermediate Transformer Representations for Enhanced Reasoning.
International Conference of Learning Representations (ICLR),
2025.
Language Decision Transformers with Exponential Tilt for Interactive Text Environments.
ArXiv,
2024.
StarCoder: may the source be with you!.
Transactions on Machine Learning Research (TMLR),
2023.
Does entity abstraction help generative Transformers reason? .
Transactions on Machine Learning Research,
2022.
Measuring Systematic Generalization in Neural Proof Generation with Transformers.
Conference on Neural Information Processing Systems (NeurIPS),
2020.