Three papers from our lab will appear at ICLR 2026:
“Robust Federated Inference” by Akash Dhasade, Sadegh Farhadkhani, Rachid Guerraoui, Nirupam Gupta, Maxime Jacovella, Anne-Marie Kermarrec, and Rafael Pinot
“Navigating the Accuracy-Size Trade-Off with Flexible Model Merging” by Akash Dhasade, Divyansh Jhunjhunwala, Milos Vujasinovic, Gauri Joshi, and Anne-Marie Kermarrec
“Optimizing Agent Planning for Security and Autonomy” by Aashish Kolluri, Rishi Sharma, Manuel Costa, Boris Köpf, Tobias Nießen, Mark Russinovich, Shruti Tople, and Santiago Zanella-Béguelin
“Leveraging Approximate Caching for Faster Retrieval-Augmented Generation” by Shai Bergman, Anne-Marie Kermarrec, Diana Petrescu, Rafael Pires, Mathis Randl, Martijn de Vos, and Ji Zhang has been accepted at Middleware 2025.
Akash Dhasade was awarded the prestigious SNSF Postdoc.Mobility Grant.
Sayan Biswas was awarded Young Scholars’ Development Program Fellowship from ACM CCS 2025.
“Robust ML Auditing using Prior Knowledge” by Jade Garcia Bourrée, Augustin Godinot, Martijn De Vos, Milos Vujasinovic, Sayan Biswas, Gilles Tredan, Erwan Le Merrer, and Anne-Marie Kermarrec has been accepted as spotlight (top 2.6% papers) at ICML 2025.
“Practical Federated Learning without a Server” by Akash Dhasade, Anne-Marie Kermarrec, Erick Lavoie, Johan Pouwelse, Rishi Sharma, and Martijn de Vos
“Leveraging Approximate Caching for Faster Retrieval-Augmented Generation” by Shai Aviram Bergman, Zhang Ji, Anne-Marie Kermarrec, Diana Petrescu, Rafael Pires, Mathis Randl, and Martijn de Vos
“Efficient Federated Search for Retrieval-Augmented Generation” by Rachid Guerraoui, Anne-Marie Kermarrec, Diana Petrescu, Rafael Pires, Mathis Randl, and Martijn de Vos
“Accelerating MoE Model Inference with Expert Sharding” by Oana Balmau, Anne-Marie Kermarrec, Rafael Pires, André Loureiro Espírito Santo, Martijn de Vos, and Milos Vujasinovic
“Harnessing Increased Client Participation with Cohort-Parallel Federated Learning” by Akash Dhasade, Anne-Marie Kermarrec, Tuan-Ahn Nguyen (Independent Researcher), Rafael Pires, and Martijn de Vos
“Low-Cost Privacy-Preserving Decentralized Learning” (in PoPETS Issue 3)by Sayan Biswas, Davide Frey, Romaric Gaudel, Anne-Marie Kermarrec, Dimitri Lerévérend, Rafael Pires, Rishi Sharma, and François Taïani
“Noiseless Privacy-Preserving Decentralized Learning” (in PoPETS Issue 1) by Sayan Biswas, Mathieu Even, Anne-Marie Kermarrec, Laurent Massoulie, Rafael Pires, Rishi Sharma, and Martijn de Vos
“Boosting Asynchronous Decentralized Learning with Model Fragmentation” by Sayan Biswas, Anne-Marie Kermarrec, Alexis Marouani, Rafael Pires, Rishi Sharma, and Martijn de Vos has been accepted at the 34th ACM Web Conference (WWW 2025) and is selected for oral presentation (top 7% papers).
Prof. Anne-Marie Kermarrec has been appointed by French President Emmanuel Macron to the newly established Presidential Science Council, joining 11 other distinguished scientists from across all disciplines.
Anastasiia Kucherenko successfully defended her PhD thesis.
“Revisiting Ensembling in One-Shot Federated Learning” by Youssef Allouah, Akash Dhasade, Rachid Guerraoui, Nirupam Gupta, Anne-Marie Kermarrec, Rafael Pinot, Rafael Pires, and Rishi Sharma was accepted at NeurIPS 2024.
“QuickDrop: Efficient Federated Unlearning via Synthetic Data Generation” by Akash Dhasade, Yaohong Ding, Song Guo, Anne-Marie Kermarrec, Martijn de Vos, and Leijie Wu was accepted at the 25th International Middleware Conference (MIDDLEWARE 2024).
“Fairness Auditing with Multi-Agent Collaboration” by Martijn de Vos, Akash Dhasade, Jade Garcia Bourrée, Anne-Marie Kermarrec, Erwan Le Merrer, Benoit Rottembourg, and Gilles Tredan was accepted at the 27th European Conference on Artificial Intelligence (ECAI 2024).
