Alexandre Piche
Research Scientist
AI Frontier Research
Alexandre Piche is a Senior Research Scientist at ServiceNow Research. He is interested in getting large language models to self-improve and self-align under minimal human supervision. Alexandre is completing his PhD at Mila, Université de Montreal and has previously interned at DeepMind, London, and Riken AIP, Tokyo (remote).
Intérêts
- Reinforcement Learning
Publications
BigCharts-R1: Enhanced Chart Reasoning with Visual Reinforcement Finetuning.
Conference on Language Modeling (COLM),
2025.
How to Train Your LLM Web Agent: A Statistical Diagnosis (Oral).
Workshop at the International Conference of Machine Learning (ICML),
2025.
AgentMerge: Enhancing Generalization in Fine-Tuned LLM Agents.
Workshop at the Neural Information Processing Systems (NeurIPS),
2024.
Fine-Tuning Web Agents: It Works, But It's Trickier Than You Think.
Workshop at the Neural Information Processing Systems (NeurIPS),
2024.
Fine-Tuning Web Agents: It Works, But It's Trickier Than You Think.
NOW AI Conference (NOWAI),
2024.
Exploring validation metrics for offline model-based optimisation with diffusion models.
Transactions on Machine Learning Research (TMLR),
2024.
Self-evaluation and self-prompting to improve the reliability of LLMs.
Workshop at the International Conference of Learning Representation (ICLR),
2024.
Bridging the Gap Between Target Networks and Functional Regularization.
Transactions on Machine Learning Research (TMLR),
2023.
Causal Discovery with Language Models as Imperfect Experts.
Workshop on Structured Probabilistic Inference & Generative Modeling (ICML),
2023.
Mastering the Unsupervised Reinforcement Learning Benchmark from Pixels.
International Conference on Machine Learning (ICML),
2023.
Leveraging Human Preferences to Master Poetry.
AAAI Workshops,
2023.
Can large language models build causal graphs?.
Workshop at the Neural Information Processing Systems (NeurIPS),
2022.
Implicit Offline Reinforcement Learning via Supervised Learning.
Workshop at the Neural Information Processing Systems (NeurIPS),
2022.
Unsupervised Model-based Pre-training for Data-efficient Reinforcement Learning from Pixels.
Workshop at the International Conference on Machine Learning (ICML),
2022.
A Probabilistic Perspective on Reinforcement Learning via Supervised Learning.
Workshop at the International Conference on Learning Representations (ICLR),
2022.
Planning with Latent SImulated Trajectories.
Workshop at the International Conference on Learning Representations (ICLR),
2019.
Probabilistic Planning with Sequential Monte Carlo Methods.
International Conference on Learning Representations (ICLR),
2019.
LLMs can learn self-restraint through iterative self-reflection.
Transactions on Machine Learning Research (TMLR),
2017.