Oxford ML Summer School 2025 // Recap 🎓
Spent a week at the Oxford ML Summer School (OxML 2025) - one of the most intense and rewarding deep dives into machine learning I’ve had in a while.
Oxford Machine Learning School 2025 is over 😔 but I’m glad to report that the quality of this summer school has been monotonically increasing 📈 I should know seeing that this is my 3rd one and probably last 😁
Days 3 and 4 brought some of the best lectures I’ve seen in this school 🗣 Ashley Edwards / Google DeepMind gave us a whirlwind view of image and video models (someone said she was presenting at 2x speed 😅 they meant it in a good way) and her work on #Genie; Yali Du / King’s College London (our program chair) talked about agent cooperation and a new paradigm for policy learning based on Understanding, Rehearsing, and Introspecting; Christian Rupprecht / University of Oxford argued that we (still) don’t have a foundation model for computer vision and pointed to what we need to finally get one; last but certainly not least, Valentina Pyatkin / Allen Institute demonstrated what it truly means to do open research on foundation models and why that matters.
🏆 As it is by now a sort of tradition, it’s time to give out some medals for the best talks: Yali takes bronze 🥉, Ashley and Christian take the 🥈 (ex aequo), and gold goes to Valentina 🥇 for the supreme importance of her work and for waking everyone up in the last session of the school. Honorable mentions go to Abdul Fatir Ansari / Amazon Web Services (AWS) and Fazl Barez / University of Oxford 📜 for delivering amazing sessions on time series and AI safety 🙏 Finally, I’d like tl thank AI for Global Goals, the organizers (Mona Alinejad, D.Phil. (Oxon) , Reza Khorshidi, D.Phil. (Oxon) and Rochelle Enriquez Radiša), the TA team (Pelin Balcı, Vlad Timu and Yitai Cheng), and all the online and onsite participants for making this event such a huge success! 🏞 #QotD comes from something I saw carved on a nearby park bench: “Learn from yesterday, live for today, hope for tomorrow”
