New workshop paper Workshop on Data-centric Machine Learning Research! Here, we intro our efforts to emulate HOD simulations (which are themselves conducted upon N-body dark matter simulations 😅) for galaxy intrinsic alignment correlations using NNs. Full journal paper + code coming soon! arxiv.org/abs/2404.13702

Will be attending ICLR 2026 and the Workshop on Data-centric Machine Learning Research in Vienna. Would love to chat about all things Alignment, Interpretability, Data-Centric AI, and how models (should) deal with conflicting training data, or any application that excites you :) DM/Reply #ICLR #ICLR2024

Reviewing recognitions are always satisfying 😌

The DMLR Workshop ICLR 2026 is over! Thank you so much to our wonderful presenters, speakers, panelists and organization team! See you at the DMLR Workshop on Datasets for Foundation Models ICML Conference in July again 🙂 Join our Discord at discord.gg/FswYXMv4j9

Excited to share our paper on "Quantifying Spuriousness of Biased Datasets Using Partial Information Decomposition" (accepted at Data-Centric Machine Learning: Datasets for Foundation Models Workshop on Data-centric Machine Learning Research at ICML Conference ) arxiv.org/abs/2407.00482 #ICML2024 #DMLR

Glad to share this paper got accepted Workshop on Data-centric Machine Learning Research ICML Conference #ICML2024 🎉 and will be presented during the workshop’s poster session on Saturday, 4pm.

CLRS-Text is incorporated in the CLRS benchmark codebase (github.com/google-deepmin…). To use it, simply `pip install dm-clrs` We also have a companion paper (arxiv.org/abs/2406.04229), to be presented Workshop on Data-centric Machine Learning Research #ICML2024 this Saturday! Please stop by if you'd like a chat 🚀

Lastly, we're delighted to unveil CLRS-Text at the Workshop on Data-centric Machine Learning Research 🚀 Larisa Markeeva Sean McLeish Avi Schwarzschild Tom Goldstein Alex Vitvitskyi & I will be around during the day and very happy to discuss it! I've already described our dataset in great detail here: x.com/PetarV_93/stat…

If you're still at ICML Conference 🇦🇹 and want to talk about Major TOM, come and see our poster at the Workshop on Data-centric Machine Learning Research to chat about open AI-ready data for Earth observation 🌍

Excited to be in Vienna for ICML Conference to present “Class-Level Code Generation from Natural Language Using Iterative, Tool-Enhanced Reasoning over Repository” (github.com/microsoft/repo…) at Workshop on Data-centric Machine Learning Research 4PM today. Excited to connect with #ML4Code & LLM Agent enthusiasts! #icml

starting Workshop on Data-centric Machine Learning Research with a talk by Aditi Raghunathan on data curation

next up Stella Biderman @ ICML on copyright

now panel on datasets for foundation models, from left Lucas Beyer (bl16) Alex Dimakis brandon Nomic AI Matthias Gerstgrasser Angéline Pouget first hot take: resnet is a foundation model (;

now Lucas Beyer (bl16) on vision x llms

next Angéline Pouget on diversity in contrastive vlms

last talk before posters Matthias Gerstgrasser asks: is model collapse inevitable? cc recent nature cover (;