Luisa Zintgraf

Huge shout-out to my co-first authors @dancalian.bsky.social, @gregfar.bsky.social, & Iurii Kemaev.

And to our amazing collaborators: Matteo Hessel, Jeremy Shar, Junhyuk Oh, András György, @schaul.bsky.social, @jeffdean.bsky.social, Hado van Hasselt, & Dave Silver.

We believe that the DataRater is a promising step towards more automated and principled dataset curation. This could be especially important for filtering and making the best use of massive synthetic datasets in the future.

For a deeper dive, check out arxiv.org/pdf/2505.17895

So what does the DataRater learn? It automatically identifies and down-weights data that aligns with human intuitions of low quality, such as incorrect text encodings, OCR errors, and irrelevant content.

The result? The DataRater is highly effective at filtering data, leading to significant compute efficiency improvements. In our experiments, we observed up to a 46.6% net compute gain while often improving final model performance.

We introduce the DataRater, a meta-learning method that learns to rate the value of each data point for training. Instead of manually specifying filtering rules, we train the DataRater to optimize for a simple goal: improving the training efficiency on a held-out dataset.

Foundation models are trained on large datasets, but not all data is created equal. Dataset curation often relies on manual, coarse-grained filtering and hand-crafted rules. This is becoming a major challenge, especially with the rise of synthetic data.

Excited to share our new paper, "DataRater: Meta-Learned Dataset Curation"!

We explore a fundamental question: How can we *automatically* learn which data is most valuable for training foundation models?

Paper: arxiv.org/pdf/2505.17895 to appear at @neuripsconf.bsky.social

Thread 👇

Tagging first author @jakeabeck.bsky.social who just joined bsky! Welcome 🎉

[AUTOML23] A Tutorial on MetaReinforcement Learning YouTube video by AutoMLConf

📘 Journal: nowpublishers.com/article/Deta...
📝 ArXiv: arxiv.org/abs/2301.08028
🎙️ Podcast: www.talkrl.com/episodes/jac...
🎥 Talk: youtu.be/XUQ9jLOZqGc

🎉 Our Meta-RL survey is now published in Foundations and Trends in Machine Learning! A deep dive into how agents can learn to learn 🤖🧠

Huge kudos to Jake Beck & Risto Vuorio for leading the charge, and to co-authors Evan Liu, Zheng Xiong, Chelsea Finn & @shimon8282.bsky.social!

Wanna work on Gemini? DeepMind is hiring! 🚀

Research Scientist, Large Scale Pre-Training Model London, UK

Interested in helping us make Gemini Pro even better?

The Gemini pre-training team is looking for a Research Scientist in London to push the boundaries of LLM scaling: understanding, predicting, and improving. ♊️🚀

Apply here: boards.greenhouse.io/deepmind/job...