Florian Le Bronnec

I am a PhD student between the teams MILES from LAMSADE (Paris-Dauphine University) and MLIA from ISIR. Don't hesitate to get in touch if we have a common research interest!

News

• ICML 2025. I attended the ICML 2025 conference in Vancouver.
• ICLR 2025. I attended the ICLR 2025 conference in Singapore.
• ACL 2024. I attended the ACL 2024 conference in Bangkok.
• Best Paper Award EACL 2024. Very proud to have been awarded the Best Paper Award for our work LOCOST: State-Space Models for Long Document Abstractive Summarization.
• EACL 2024. I attended the EACL 2024 conference in Malta.

Research Interests

• Probabilist machine learning
• Large Language Models
• Efficient generative architectures
• Evaluation of Large Language Models
• Faithfulness of LLMs

Short Bio

• 2023-Today: PhD student in Machine Learning for NLP, at Universite Paris Dauphine-PSL. I'm part of the labs LAMSADE (MILES Team) and ISIR (MLIA Team)
• May 2022-October 2022: SIPGA Awardee. Research intern at A*STAR (I2R) in Singapore under the supervision of Nancy F. Chen and Mathieu Ravaut. Worked on long documents abstractive summarization.
• 2021–2022: MSc in Applied Mathematics at École Normale Supérieure Paris-Saclay.
• 2018–2022: Engineering degree (BSc + MSc equivalent) from CentraleSupélec, Université Paris-Saclay. Specialized in mathematics with a focus on machine learning and statistics.

A Few Recent Publications

• Florian Le Bronnec*, Alexandre Vérine*, Benjamin Negrevergne, Yann Chevaleyre, Alexandre Allauzen, Improving Diversity in Language Models: When Temperature Fails, Change the Loss (ICML 2025)

• Florian Le Bronnec*, Song Duong*, Alexandre Allauzen, Laure Soulier, Vincent Guigue, Alberto Lumbreras, Patrick Gallinari SCOPE: A Self-supervised Framework for Improving Faithfulness in Conditional Text Generation (ICLR 2025)

• Florian Le Bronnec*, Alexandre Verine*, Benjamin Negrevergne, Yann Chevaleyre, Alexandre Allauzen Exploring Precision and Recall to assess the quality and diversity of LLMs (ACL 2024)

• Florian Le Bronnec*, Song Duong*, Mathieu Ravaut, Alexandre Allauzen, Nancy F. Chen, Vincent Guigue, Alberto Lumbreras, Laure Soulier, Patrick Gallinari LOCOST: State-Space Models for Long Document Abstractive Summarization (EACL 2024, Best Paper Award)

Current Teaching Activities

• Large Language Models (M2), M2 IASD
• Natural Langauge Processing (Executive Master), Exec Master IASD

Links

• Google Scholar
• GitHub

Contact Information

• e-mail: florian.le-bronnec (at) dauphine.psl.eu