Our paper "Collaborative Classification from Noisy Labels" has been accepted to #AISTATS2021. It introduces a Bayesian model that fixes errors in metadata by taking advantage of user-item interactions. With Lucas Maystre, N Kumarappan & J Bütepage. Final version to come soon!

Looking forward to present our paper at AISTATS this afternoon. Hope to see you there!

Tomorrow @fraspass will present our work on modeling the long-term evolution of user preferences at #TheWebConf . Paper will be available soon, in the meantime check out our blog post!

Happy to share that our paper "The Dynamics of Exploration on Spotify" was accepted for #ICWSM22. Work led by Spotify Research intern, Lillio Mok, with Sam Way, Lucas Maystre, and Ashton Anderson. Congrats, Lillio and team!

We look forward to presenting our paper Active Offline Policy Selection at #NeurIPS2021! paper: openreview.net/pdf?id=Zsrn9wX… code: github.com/deepmind/activ… website: sites.google.com/corp/view/acti… Yutian Chen Tom Le Paine Caglar Gulcehre Cosmin Paduraru Daniel J. Mankowitz Misha Denil Nando de Freitas

New work shows that active offline policy selection (A-OPS) can accelerate policy development in real-world applications like robotics. A-OPS helps to quickly identify the best policy even when evaluation time on the robot is very limited: dpmd.ai/AOPS 1/

Also happy to announce that "Multistate Analysis with Infinite Mixtures of Markov Chains" has also been accepted to #UAI2022. Work by Lucas Maystre, Tiffany Wu, Roberto Sanchis & Tony Jebara

We work on cutting-edge research in Machine Learning, so we are excited to be sponsoring this year’s NeurIPS conference and the WiML workshop. If you are attending #NeurIPS2022 in New Orleans, please come by the Spotify booth and say hello! research.atspotify.com/2022/11/spotif…

What ideas can RL bring to survival analysis? And vice-versa? Daniel Russo and I explore this in our #NeurIPS2022 paper. Check out our post on the Spotify Research blog to get a quick overview of our work. research.atspotify.com/2022/11/surviv…

Are you a PhD student interested in tackling real-world problems using ML? My team (Spotify Research) is hiring interns for next summer. lifeatspotify.com/jobs/summer-in…

How can we develop more generalisable reward models for agent behaviours? Excited to share my @deepmind internship project, where we investigate finetuning Flamingo🦩w/ human reward annotations to train success detectors in 3 different domains! 📜arxiv.org/abs/2303.07280 🧵1/

Are your perceptions of the #carbon emissions you generate on a daily basis aligned with reality? Take the Climpact test to find out! Victor Kristof Lucas Maystre #CO2 #ClimateAction EPFL actu.epfl.ch/news/climpact-…

How can we optimize recommendations for long-term user satisfaction, but avoid cold-start for new content? We address this in our #KDD2023 paper: "Impatient Bandits: Optimizing for the Long-Term Without Delay" Work by @tomcd_ Lucas Maystre Mounia Lalmas Daniel Russo & Kamil Ciosek.

How can we optimize recommendations for long-term user satisfaction, but avoid cold-start for new content? You can watch a short video of our upcoming #KDD2023 paper addressing this on youtube.com/watch?v=yYdBru… … and check our blog post Spotify Research. research.atspotify.com/2023/07/optimi…

Our upcoming #KDD2023 paper presents our journey on prototyping Reinforcement Learning based playlist personalization at Spotify. You can watch a short video of our work at youtube.com/watch?v=Ysaa8O… and read our latest blog post Spotify Research. research.atspotify.com/2023/07/automa…

We are looking for research scientist interns in the fields of AI, ML, economics, LLMs, recommender systems, speech processing, NLP, or IR. These internships are based in our London office. Apply at lifeatspotify.com/jobs/2024-summ…

🚀Announcing the Workshop on Computer Use Agents at #ICML2025 in July, Vancouver! Join us, to advance research on AI agents performing real-world computer tasks. 🤖Call for Papers & Demos: Deadline May 18, 2025 🎙️Exciting speaker lineup announced! ✍️Interested in

Meet Gemini Robotics 1.5! Our ER model brings embodied reasoning to the real world. It understands high-level goals like "pack ingredients for mushroom risotto" with planning & success detection. Also, check out the cool "active vision" behavior – observing the action up close!🔍

Interacting with robots just got easier with the agentic capabilities of Gemini Robotics 1.5. Talk to the robot or show it things! See how the ER model reads a handwritten list on paper and packs the tools for the job.

An update: I have left Meta Superintelligence Labs and joined Reflection AI in NYC!! Today is my first day. I started in the Fundamental AI Research (FAIR) lab at Meta, then Facebook, over six (!) years ago as my first job out of the PhD. They were some formative years. The