Publications
You can also find my articles on my Google Scholar profile.
2026
-
Flower Hub: A Reproducible Benchmarking Platform for Federated Learning in Simulation and Deployment
Flower Hub is a platform for publishing, discovering, and executing decentralized and federated applications, enabling reproducible benchmarking across simulation and deployment.
-
The Red Queen Gödel Machine: Co-Evolving Agents and Their Evaluators
A self-improving system in which agents and the evaluators that score them co-evolve, avoiding the reward hacking that fixed evaluators invite.
-
FoMoE: Breaking the Full-Replica Barrier with a Federation of MoEs
FoMoE breaks the full-replica barrier in distributed training by federating mixtures of experts, so no participant needs to hold the entire model.
-
Response-Conditioned Parallel-to-Sequential Orchestration for Multi-Agent Systems
A multi-agent orchestration scheme that adaptively transitions between parallel and sequential execution based on intermediate agent responses.
-
LoRDO: Distributed Low-Rank Optimization with Infrequent Communication
LoRDO performs distributed low-rank optimization with infrequent communication, matching full-rank quality at a fraction of the synchronisation cost.
2025
-
MT-DAO: Multi-Timescale Distributed Adaptive Optimizers with Local Updates
MT-DAO combines optimizer moments tracked on multiple timescales with local updates, closing the gap to fully synchronous training under infrequent communication.
-
AbbIE: Autoregressive Block-Based Iterative Encoder for Efficient Sequence Modeling
AbbIE is an autoregressive block-based iterative encoder that recursively refines representations, improving perplexity at a fixed parameter budget.
-
DES-LOC: Desynced Low Communication Adaptive Optimizers for Training Foundation Models
DES-LOC desynchronises the communication of optimizer states, cutting bandwidth for foundation-model training without sacrificing convergence.
-
LUNAR: LLM Unlearning via Neural Activation Redirection
We present LUNAR, a novel unlearning methodology grounded in the Linear Representation Hypothesis.
-
SparsyFed: Sparse Adaptive Federated Training
This work proposes a novel approach to sparse federated training that does not require to fix a global mask to be efficient.
2024
-
Photon: Federated LLM Pre-Training
We present Photon, the first fully federated system for the federated pre-training of large language models.
-
DEPT: Decoupled Embeddings for Pre-training Language Models.
This work proposes a novel pre-training method for language models that decouples the embeddings from the rest of the model.
-
Sheaf HyperNetworks for Personalized Federated Learning
Sheaf HyperNetworks is a novel HyperNetwork-based methodology to improve the performance of any personalized federated learning setting.
-
Worldwide Federated Training of Language Models
WorldLM is a novel approach to federated training of language models exploiting federations of federations in a hierarchical setting.
-
The Future of Large Language Model Pre-training is Federated
We present in this work the first fully federated system for the federated pre-training of large language models.
-
FedAnchor: Enhancing Federated Semi-Supervised Learning with Label Contrastive Loss for Unlabeled Clients
FedAnchor is a methodology to solve federated semi-supervised learning problems.
2023
-
High-throughput Simulation of Federated Learning via Resource-Aware Client Placement
This paper is about Pollen, a system that enables high-throughput simulation of federated learning.
2022
-
Flower: A Friendly Federated Learning Research Framework
This paper is about Flower, the federated learning framework developed at the University of Cambridge by the MLSys group.