Nicolas Usunier

Researcher, H Company

Joined

September 2016

Names

Nicolas Usunier (Preferred)

Emails

****@fb.com (Confirmed)

****@meta.com (Confirmed)

****@gmail.com (Confirmed)

****@hcompany.ai (Confirmed)

Personal Links

Google Scholar

DBLP

Career & Education History

Researcher

H Company (hcompnay.ai)

2024 – Present

Research Scientist

Facebook (fb.com)

2015 – 2024

Advisors, Relations & Conflicts

No relations added

Expertise

ranking and preference learning

matrix and tensor factorization

learning theory

structured prediction

supervised learning

Present

Publications

Winter Soldier: Backdooring Language Models at Pre-Training with Indirect Data Poisoning
Wassim Bouaziz, Mathurin VIDEAU, Nicolas Usunier, El-Mahdi El-Mhamdi
- ICLR 2026 Poster
- Readers: Everyone
Winter Soldier: Backdooring Language Models at Pre-Training with Indirect Data Poisoning
Wassim Bouaziz, Mathurin VIDEAU, Nicolas Usunier, El-Mahdi El-Mhamdi
- NeurIPS Lock-LLM Workshop 2025 Poster
- Readers: Everyone
Inverting Gradient Attacks Makes Powerful Data Poisoning
Wassim Bouaziz, Nicolas Usunier, El-Mahdi El-Mhamdi
- Accepted by TMLR
- Readers: Everyone
WINTER SOLDIER: HYPNOTIZING LANGUAGE MODELS AT PRE-TRAINING WITH INDIRECT DATA POISONING
Wassim Bouaziz, Mathurin VIDEAU, Nicolas Usunier, El-Mahdi El-Mhamdi
- WMARK@ICLR2025
- Readers: Everyone
Group fairness without demographics using social networks
David Mingfei Liu, Virginie Do, Nicolas Usunier, Maximilian Nickel
- FAccT
- Readers: Everyone
Data Taggants: Dataset Ownership Verification Via Harmless Targeted Data Poisoning
Wassim Bouaziz, Nicolas Usunier, El-Mahdi El-Mhamdi
- ICLR 2025 Poster
- Readers: Everyone
Code Llama: Open Foundation Models for Code
Jingyu Liu, Baptiste Roziere, Jonas Gehring, Fabian Gloeckle, Sten Sootla, Itai Gat, Xiaoqing Tan, Yossi Adi, Tal Remez, Jérémy Rapin, Ivan Evtimov, Joanna Bitton, Manish Bhatt, Cristian Canton Ferrer, Wenhan Xiong, Alexandre Défossez, Jade Copet, Hugo Touvron, Louis Martin, Nicolas Usunier et al. (2 additional authors not shown)
- OpenReview Archive Direct Upload
- Readers: Everyone
From gradient attacks to data poisoning
Wassim Bouaziz, El-Mahdi El-Mhamdi, Nicolas Usunier
- ICLR 2024 Conference Withdrawn Submission
- Readers: Everyone
Towards Reliable Assessments of Demographic Disparities in Multi-Label Image Classifiers
Melissa Hall, Bobbie Chern, Laura Gustafson, Denisse Ventura, Harshad Kulkarni, Candace Ross, Nicolas Usunier
- 2023 (modified: 06 Mar 2023)
- CoRR 2023
- Readers: Everyone
Leveraging Demonstrations with Latent Space Priors
Jonas Gehring, Deepak Gopinath, Jungdam Won, Andreas Krause, Gabriel Synnaeve, Nicolas Usunier
- Accepted by TMLR
- Readers: Everyone

View all 141 publications

Co-Authors

View all 156 co-authors