Nicolas Bougie

Nicolas Bougie

Research Scientist

Woven by Toyota

Welcome!

I am a Research Scientist at Woven by Toyota, where I work on building intelligent systems that understand and simulate human behavior. My research focuses on developing AI agents that can faithfully reproduce human decision-making patterns, with applications in recommender systems, autonomous systems, and human-AI interaction.

My work bridges the gap between large language models (LLMs) and real-world user behavior. I am particularly interested in how we can train AI agents to internalize world dynamics and align with human preferences through counterfactual reasoning, self-reflection, and intrinsic motivation.

Prior to Woven by Toyota, I worked on reinforcement learning and intrinsic motivation, exploring how agents can learn efficiently through curiosity-driven exploration and hierarchical learning from human feedback.

Research Interests

LLM Agents User Simulation Recommender Systems World Models Reinforcement Learning Intrinsic Motivation Human-AI Alignment

Recent News

Jan 2025
AlignUSER

AlignUSER: Human-Aligned LLM Agents via World Models

New paper introducing world-model-driven agents that achieve 52.9% action prediction accuracy (vs. 24.2% prior best) for recommender system evaluation.

Read more →
2026
MobileCity

MobileCity accepted to EACL 2026

Our paper on efficient large-scale urban behavior simulation with LLM agents has been accepted to EACL 2026.

Read more →
2025
SimUSER

SimUSER accepted to ACL 2025

Our paper on simulating user behavior with large language models for recommender system evaluation has been accepted to ACL 2025.

Read more →
2025
CitySim

CitySim accepted to EMNLP 2025

CitySim presents large-scale LLM-driven agent simulation for modeling urban behaviors and city dynamics.

Read more →

Selected Publications Google Scholar

View all
AlignUSER
arXiv 2025

AlignUSER: Human-Aligned LLM Agents via World Models for Recommender System Evaluation 🔥

Nicolas Bougie, Gian Marconi Marconi, Tony Yip, Narimasa Watanabe

arXiv preprint, 2025

SimUSER
ACL 2025

SimUSER: Simulating User Behavior with Large Language Models for Recommender System Evaluation

Nicolas Bougie, Narimasa Watanabe

Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (ACL), 2025

MobileCity
EACL 2026

MobileCity: An Efficient Framework for Large-Scale Urban Behavior Simulation

Xiaotong Ye*, Nicolas Bougie*, Toshihiko Yamasaki, Narimasa Watanabe

Proceedings of the 2026 Conference of the European Chapter of the Association for Computational Linguistics (EACL), 2026

CitySim
EMNLP 2025

CitySim: Modeling Urban Behaviors and City Dynamics with Large-Scale LLM-Driven Agent Simulation

Nicolas Bougie, Narimasa Watanabe

Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing (EMNLP), 2025

Generative Reviewer Agents
ACL 2025

Generative Reviewer Agents: Scalable Simulacra of Peer Review

Nicolas Bougie, Narimasa Watanabe

Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (ACL), 2025

Fast and Slow Curiosity
Applied Intelligence 2021

Fast and Slow Curiosity for High-Level Exploration in Reinforcement Learning

Nicolas Bougie, Ryutaro Ichise

Applied Intelligence, Springer, 2021

Skill-based Curiosity
Machine Learning 2020

Skill-Based Curiosity for Intrinsically Motivated Reinforcement Learning

Nicolas Bougie, Ryutaro Ichise

Machine Learning, vol. 109, pp. 493-512, Springer, 2020

Progress-Driven Exploration
ICANN 2020 Best Paper

Exploration via Progress-Driven Intrinsic Rewards

Nicolas Bougie, Ryutaro Ichise

International Conference on Artificial Neural Networks (ICANN), 2020

Interpretable Reinforcement Learning
ACM TIST 2024

Interpretable Reinforcement Learning with Neural Symbolic Logic

Nicolas Bougie, Ryutaro Ichise

ACM Transactions on Intelligent Systems and Technology (TIST), 2024