Nicholas Roberts
@nick11roberts.bsky.social
Ph.D. student at UW-Madison. Working on automating foundation model guided science. Previously at CMU, UCSD, Fresno City College.
https://nick11roberts.science
https://nick11roberts.science
Reposted by Nicholas Roberts
Online data mixing reduces training costs for foundation models, but faces challenges:
⚠️ Human-defined domains miss semantic nuances
⚠️ Limited eval accessibility
⚠️ Poor scalability
Introducing 🎵R&B: first regroup data, then dynamically reweight domains during training!
⚠️ Human-defined domains miss semantic nuances
⚠️ Limited eval accessibility
⚠️ Poor scalability
Introducing 🎵R&B: first regroup data, then dynamically reweight domains during training!
May 8, 2025 at 5:01 PM
Online data mixing reduces training costs for foundation models, but faces challenges:
⚠️ Human-defined domains miss semantic nuances
⚠️ Limited eval accessibility
⚠️ Poor scalability
Introducing 🎵R&B: first regroup data, then dynamically reweight domains during training!
⚠️ Human-defined domains miss semantic nuances
⚠️ Limited eval accessibility
⚠️ Poor scalability
Introducing 🎵R&B: first regroup data, then dynamically reweight domains during training!
Reposted by Nicholas Roberts
Today at @iclr-conf.bsky.social, come chat with @changho.bsky.social about what types of data drive weak-to-strong generalization!
April 23, 2025 at 8:29 PM
Today at @iclr-conf.bsky.social, come chat with @changho.bsky.social about what types of data drive weak-to-strong generalization!
Reposted by Nicholas Roberts
First up at #NeurIPS2024 from our group, our work on labeling via programmatic distillation (a spotlight!). Label your data orders of magnitude faster and cheaper — come join us today at Poster Session 2 East for a demo!
December 11, 2024 at 11:15 PM
First up at #NeurIPS2024 from our group, our work on labeling via programmatic distillation (a spotlight!). Label your data orders of magnitude faster and cheaper — come join us today at Poster Session 2 East for a demo!