Fred Sala
@fredsala
Assistant Professor @WisconsinCS. Chief scientist @SnorkelAI. Working on machine learning & information theory.
Happy to see a time series workshop @ NeurIPS 2025 motivated in part by our search for BERT moments in specialized foundation models (🧵here: x.com/khodakmoments/…)
🚀 We are happy to organize the BERT²S workshop @NeurIPSConf 2025 on Recent Advances in Time Series Foundation Models. 🌐 berts-workshop.github.io 📜Submit by August 22 🎓Speakers and panelists: @ChenghaoLiu15 Mingsheng Long @zoe_piran @danielle_maddix @atalwalkar @qingsongedu
I'll be traveling to ICML to present our work on data mixtures at two workshops on Saturday (DataWorld + DIG-BUGS). looking forward to attending my first in-person conference and connecting with others!
Excited to be at #ICML2025. Join us at poster session today at 11am PDT Poster no. E-1304 (East Hall) to check out our new lens on SSL: Rethinking Confidence Scores and Thresholds in Pseudolabeling-based SSL openreview.net/pdf?id=w4c5bLk… @harit_v @kitkatdafu @srinath_namburi @fredsala
Heading to #ICML! I’ll be representing SprocketLab at @UWMadison and @SnorkelAI. Reach out if you want to chat about data-centric AI, data development, agents, and foundation models.
Efficient data curation is critical for modern ML. 📣 We introduce Mimic Score, a new, lightweight, model-based metric for sample utility that leverages reference model's weights to identify high-value samples and accelerate training. 🎉 Accepted as an Oral at ICML’25 DataWorld!
Very exciting work on using weak supervision for RL- closing the “generation-verification gap”!! Once again- principled approaches to labeling/data development are the keys!
How can we close the generation-verification gap when LLMs produce correct answers but fail to select them? 🧵 Introducing Weaver: a framework that combines multiple weak verifiers (reward models + LM judges) to achieve o3-mini-level accuracy with much cheaper non-reasoning…
🚨Test-time scaling for code has a MASSIVE scalability issue on the horizon: the time to verify a candidate. ✅ORMs are a scalable solution through a generate-prune-then-rank approach. ✅This can be 11.65x faster than the full verification while only being 8.33% less accurate
Reinforcement learning has driven impressive gains in LLM reasoning—but what exactly does RL improve? SPARKLE answers this question with a fine-grained evaluation framework that dissects reasoning into plan-following, problem decomposition, and knowledge use. The results are…
🧠 New research reveals why Reinforcement Learning makes language models better at reasoning (spoiler: it's not what we thought!) 📝 Paper: arxiv.org/pdf/2506.04723 💻 Website: sparkle-reasoning.github.io We performed multi-stage curriculum-style RL training from base LLMs. Key…
Scale alone is not enough for AI data. Quality and complexity are equally critical. Excited to support all of these for LLM developers with @SnorkelAI Data-as-a-Service, and to share our new leaderboard! — Our decade-plus of research and work in AI data has a simple point:…
Agentic AI will transform every enterprise–but only if agents are trusted experts. The key: Evaluation & tuning on specialized, expert data. I’m excited to announce two new products to support this–@SnorkelAI Evaluate & Expert Data-as-a-Service–along w/ our $100M Series D! ---…
I'm joining @WisconsinCS @uwcdis as an assistant professor in fall 2026!! There, I'll continue working on language models, computational social science, & responsible AI. 🌲🧀🚣🏻♀️ Apply to be my PhD student! Before then, I'll postdoc for a year at another UW🏔️ -- @uwnlp @uwcse.
If you're are ICLR, please check out @yingfan_bot's work on looped Transformers!
Excited to share that our paper Looped Transformers for Length Generalization (arxiv.org/abs/2409.15647) has been accepted at #ICLR2025! 🎉 Feel free to stop at Poster Session 3 (Fri 25 Apr 10 a.m. +08 — 12:30 p.m. +08) in Hall 3 + Hall 2B #475
🎉 Excited to present our #ICLR2025 work—leveraging future medical outcomes to improve pretraining for prognostic vision models. 🖼️ "Time-to-Event Pretraining for 3D Medical Imaging" 👉 Hall 3+2B #23 📍 Sat 26 Apr, 10 AM–12:30 PM 🔗 iclr.cc/virtual/2025/p…
Want to align LLMs and text-to-image models to heterogeneous preferences? Check out our #ICLR2025 paper on learning personalizable rewards (PAL) in a sample efficient way. iclr.cc/virtual/2025/p… When: Friday Apr 25th from 10 am Where: Hall 3 + Hall 2B #196, Poster Session 3
Today at #ICLR2025---come chat with @Changho_Shin_ about our work on what types of data drive weak-to-strong generalization!
