Publications

Fast but False Progress on Benchmarks with Richer Feedback

Y. Allouah, J. Duchi, S. Koyejo.

website·code

How Reusable Are Benchmarks with Richer Feedback?

Y. Allouah, J. Duchi.

paper·code

The Distillation Game: Adaptive Evaluations & Efficient Defenses

Y. Allouah*, M. Haghifam*, S. Koyejo, R. Shokri. (* equal contribution)

ICML Workshop on Foundations of Deep Generative Models (FoGen), 2026

paper·code

What to Forget in Unlearning? Forget Set Curation for Language Models

A. Jha, A. Khatua, Y. Allouah, S. Koyejo.

Conference on Neural Information Processing Systems (NeurIPS), 2026

paper·code

Efficient Prediction of Pass@k Scaling in Large Language Models

J. Kazdan, R. Schaeffer, Y. Allouah, C. Sullivan, K. Yu, N. Levi, S. Koyejo.

Conference on Neural Information Processing Systems (NeurIPS), 2026

paper

Distributional Machine Unlearning via Selective Data Removal

Y. Allouah, R. Guerraoui, S. Koyejo.

International Conference on Learning Representations (ICLR), 2026

ICML Workshop on Machine Unlearning for Generative AI (MUGen), 2025 (Oral)

paper·code

Certified Unlearning for Neural Networks

A. Koloskova*, Y. Allouah*, A. Jha, R. Guerraoui, S. Koyejo. (* equal contribution)

International Conference on Machine Learning (ICML), 2025

paper·code

Towards Trustworthy Federated Learning with Untrusted Participants

Y. Allouah, R. Guerraoui, J. Stephan.

International Conference on Machine Learning (ICML), 2025

paper

The Utility and Complexity of In- and Out-of-Distribution Machine Unlearning

Y. Allouah, J. Kazdan, R. Guerraoui, S. Koyejo.

International Conference on Learning Representations (ICLR), 2025

paper·slides

Adaptive Gradient Clipping for Robust Federated Learning

Y. Allouah, R. Guerraoui, N. Gupta, A. Jellouli, G. Rizk, J. Stephan.

International Conference on Learning Representations (ICLR), 2025 (Spotlight)

paper

The Privacy Power of Correlated Noise in Decentralized Learning

Y. Allouah, A. Koloskova, A. El Firdoussi, M. Jaggi, R. Guerraoui.

International Conference on Machine Learning (ICML), 2024

paper·code

Robust Sparse Voting

Y. Allouah, R. Guerraoui, L. Hoang, O. Villemaud.

International Conference on Artificial Intelligence and Statistics (AISTATS), 2024

paper·code

Revisiting Ensembling in One-Shot Federated Learning

Y. Allouah, A. Dhasade, R. Guerraoui, N. Gupta, A.-M. Kermarrec, R. Pinot, R. Pires, R. Sharma.

Conference on Neural Information Processing Systems (NeurIPS), 2024

paper

Fine-Tuning Personalization in Federated Learning to Mitigate Adversarial Clients

Y. Allouah, A. El Mrini, R. Guerraoui, N. Gupta, R. Pinot.

Conference on Neural Information Processing Systems (NeurIPS), 2024

paper

Byzantine-Robust Federated Learning: Impact of Client Subsampling and Local Updates

Y. Allouah, S. Farhadkhani, R. Guerraoui, N. Gupta, R. Pinot, G. Rizk, S. Voitovych.

International Conference on Machine Learning (ICML), 2024

paper

Robust Distributed Learning: Tight Error Bounds and Breakdown Point under Data Heterogeneity

Y. Allouah, R. Guerraoui, N. Gupta, R. Pinot, G. Rizk.

Conference on Neural Information Processing Systems (NeurIPS), 2023 (Spotlight)

paper·code·video

On the Privacy-Robustness-Utility Trilemma in Distributed Learning

Y. Allouah, R. Guerraoui, N. Gupta, R. Pinot, J. Stephan.

International Conference on Machine Learning (ICML), 2023

paper·video

Fixing by Mixing: a Recipe for Optimal Byzantine ML under Heterogeneity

Y. Allouah, S. Farhadkhani, R. Guerraoui, N. Gupta, R. Pinot, J. Stephan.

International Conference on Artificial Intelligence and Statistics (AISTATS), 2023

paper

Further Results on Latent Discourse Models and Word Embeddings

S. Khalife, D. Goncalves, Y. Allouah, L. Liberti.

Journal of Machine Learning Research (JMLR), 2021

paper

The Cost of Trust in Machine Learning: Privacy, Robustness, Unlearning, and their Interactions

Y. Allouah.

Doctoral thesis (EPFL), 2025

thesis