Do not miss a series of deep technical contributions from the @Cambridge_Uni ML Systems lab (@CaMLSys) @iclr_conf this week in Rio. 4 conference papers, 2 workshop papers, 1 talk. I'll be there from Friday through Sunday. I'm also looking to hire current and future superstars for our frontier AI team @flwrlabs (careers.flower.ai/) and a new postdoc at CaMLSys. DM me to meet up. [paper] MT-DAO: Multi-Timescale Distributed Adaptive Optimizers - Pavilion 3 Poster Location: P3-#405 - Sat, Apr 25 2026; 4:15 PM – 6:45 PM - openreview.net/forum?id=5yPP… [paper] DES-LOC: Desynced Low Communication Adaptive Optimizers for Foundation Models - Pavilion 3 Poster Location: P3-#406 - Sat, Apr 25 2026; 4:15 PM – 6:45 PM - openreview.net/forum?id=6N2q… [paper] Cascadia: An Efficient Cascade Serving System for Large Language Models - Pavilion 3 Poster Location: P3-#1625 - Sat, Apr 25 2026; 4:15 PM – 6:45 PM - openreview.net/forum?id=wkrW… [paper] Rethinking Data Curation in LLM Training: Online Reweighting Offers Better Generalization than Offline Methods - Pavilion 3 Poster Location: P3-#519 - Thu, Apr 23 2026; 11:30 AM – 2:00 PM - (also appearing at ICLR 2026 Workshop on Navigating and Addressing Data Problems for Foundation Models Sun, Apr 26, 2026) [workshop] Homophily as a Lossy Channel: Decomposing Information in Graphs and Graph Neural Networks - Workshop on Scientific Methods for Understanding Deep Learning - Sun, April 26 2026 in Room 101-B - openreview.net/forum?id=wFBR… [talk] I don't know the title yet - Sun, April 26 2026 10.15am at Laguna Barre Hotel - informal workshop on Protocol Learning: Decentralized Collaborative Learning at Scale by @Pluralis @Alex__Iacob, @itsmaddox_j, @mher_safaryan, @meghdadkurmanji, @lorenzosani97, @sam_hrvth, @williamfshen, @xinchiqiu, @Renee42581826, @yihong_thu, Yuzhi Tang, Wentao Ma, Shengchao Hu, @shelling343, Abhinav Mehrotra, @parisgiampouras, @P_Aleks24, Youhe Jiang, Fangcheng Fu, @steverab, Jintao Zhang, Binhang Yuan, Vivek Kothari, Nicholas D. Lane University of Cambridge + @Cambridge_CL

Great to be representing @flwrlabs @italiantechweek in Turin. So many discussion about federated and privacy in AI. So many Flower users. Change is her. Ping me to meet.

Our next talk in the @CaMLSys Seminar Series is this Monday. Join us at 11am Aug 4th to hear from Salma Kharrat -- @KAUST, who will give a talk on ML under various forms of constraints -- including examples such as federatations and LLMs. "Learning Under Constraints: From Federated Collaboration to Black-Box LLMs" -- talks.cam.ac.uk/talk/index/2… Salma Kharrat (KAUST) Monday 04 August 2025, 11:00-12:00 Computer Lab, FW26. Abstract: In both federated learning (FL) and large language model (LLMs) optimization, a central challenge is effective learning under constraints, ranging from data heterogeneity and personalization to limited communication and black-box access. In this talk, I present three approaches that address these challenges across different settings. FilFL improves generalization in FL by filtering clients based on their joint contribution to global performance. DPFL tackles decentralized personalization by learning asymmetric collaboration graphs under strict resource budgets. Moving beyond FL, I will present ACING , a reinforcement learning method for optimizing instructions in black-box LLMs under strict query budgets, where weights and gradients are inaccessible. While these works tackle distinct problems, they are unified by a common goal: developing efficient learning mechanisms that perform reliably under real-world constraints.

Popularity of @flwrlabs is getting a bit out of hand in Cambridge

If you are out there auditing AI systems for compliance with AI regulations, please ping me or @billymarino -- we'd love to chat as part of our research into computational compliance being done @CaMLSys. Such discussions would inform an upcoming research paper for which we are hoping to briefly interview AI compliance professionals who are actively doing audits of AI systems and models for compliance with AI regulations like the EU AI Act (e.g., third party conformity assessments under Article 43). We would greatly appreciate any leads to potential interviewees --- or, if you might have people like this in your network, please like or reshare this post to extend its reach. More details: We can keep this anonymous or, conversely, acknowledge you in the final paper if we end up drawing on your insights. Also, there is no need to discuss details (e.g., about clients): the focus is more on the high-level features of the audit process: inputs, outputs, workflow, roles.

The Cambridge ML Systems Lab (@CaMLSys), working together with @flwrlabs, has been selected as one of a handful of teams from around the world to invent open-source decentralized ML solutions in a program funded by the German government. This farsighted initiative developed by the @SPRIND think tank could not have been more timely -- I believe the solutions we will develop will be key in maintaining an open AI landscape accessible to all. Thank you SPRIND! Our team will be co-led by @meghdadkurmanji

Four papers, one award and one keynote at @NeurIPSConf 2024 this week for the Cambridge Machine Learning Systems Lab (@CaMLSys) and our collaborators! "CLUES: Collaborative Private-domain High-quality Data Selection for LLMs via Training Dynamics" Dec 12, 00:30 GMT -- NeurIPS Conference Poster Session 2 East Paper: openreview.net/forum?id=OU1u… "Worldwide Federated Training of Language Models" FL@FM-NeurIPS'24 Outstanding Paper Award Dec 15 -- FL@FM-NeurIPS'24 Workshop (East Meeting Room 8, 15) Paper: arxiv.org/abs/2405.14446 "The Future of Large Language Model Pre-training is Federated" Dec 15 -- FL@FM-NeurIPS'24 Workshop Paper: arxiv.org/abs/2405.10853 "The First AGI will be Federated" FL@FM-NeurIPS'24 Keynote Talk Dec 15 -- FL@FM-NeurIPS'24 Workshop "Compliance Cards: Automated EU AI Act Compliance Analyses amidst a Complex AI Supply Chain" Selected for RegML Workshop Oral Presentation Dec 15 -- Regulatable ML Workshop (East Meeting Room 13) Paper: arxiv.org/abs/2406.14758 Do reach out to @CaMLSys or @flwrlabs members attending #NeurIPS2024: Wanru Zhao (@Renee42581826), Bill Marino (@billymarino), Alex Iacob (@Alex__Iacob), Lorenzo Sani (@lorenzosani97), Charles Beauville