Audrey Huang
@auddery
Followers
151
Following
88
Media
0
Statuses
20
Never too early for neurips hype
Announcing the first workshop on Foundations of Language Model Reasoning (FoRLM) at NeurIPS 2025! 📝Soliciting abstracts that advance foundational understanding of reasoning in language models, from theoretical analyses to rigorous empirical studies. 📆 Deadline: Sept 3, 2025
0
5
37
Announcing the first workshop on Foundations of Language Model Reasoning (FoRLM) at NeurIPS 2025! 📝Soliciting abstracts that advance foundational understanding of reasoning in language models, from theoretical analyses to rigorous empirical studies. 📆 Deadline: Sept 3, 2025
1
27
162
Life update!! 📣🎉 I defended my PhD thesis! Big thanks to my wonderful advisor Jeff Schneider and thesis committee @zicokolter, Aarti Singh, and Jasper Snoek. Next up: I'm joining @MicrosoftAI as a Member of Technical Staff - stoked to be back in the Bay Area!! Idemooooo 🔥💪
25
7
199
like everyone else i am hopping on the blog post trend https://t.co/t3Rma35FWC
gene.ttic.edu
A personal website.
0
30
181
I have been waiting 6 weeks for you to come and schedule us for the blind installation. We have spent 20 hours on phone on hold. Never using home depot again. @HomeDepot @RFInstallations
4
1
13
missing ICML, and I used this week to write my first technical blog on some recent thoughts on two different roles of simulators in RL and the confusions/misconceptions around them. Comments welcome! https://t.co/EAM3fIzT4g
4
19
144
Announcing the first workshop on Foundations of Post-Training (FoPT) at COLT 2025! 📝 Soliciting abstracts/posters exploring theoretical & practical aspects of post-training and RL with language models! │ 🗓️ Deadline: May 19, 2025
1
32
86
RL and post-training play a central role in giving language models advanced reasoning capabilities, but many algorithmic and scientific questions remain unanswered. Join us at FoPT @ COLT '25 to explore pressing emerging challenges and opportunities for theory to bring clarity.
Announcing the first workshop on Foundations of Post-Training (FoPT) at COLT 2025! 📝 Soliciting abstracts/posters exploring theoretical & practical aspects of post-training and RL with language models! │ 🗓️ Deadline: May 19, 2025
1
7
64
Is Best-of-N really the best we can do for language model inference? New algo & paper: 🚨InferenceTimePessimism🚨 Led by the amazing Audrey Huang (@auddery) with Adam Block, Qinghua Liu, Nan Jiang (@nanjiang_cs), and Akshay Krishnamurthy. Appearing at ICML '25. 1/11
2
23
193
Akshay presenting InferenceTimePessimism, a new alternative to BoN sampling for scaling test-time compute. From our recent paper here:
arxiv.org
Inference-time computation offers a powerful axis for scaling the performance of language models. However, naively increasing computation in techniques like Best-of-N sampling can lead to...
#AISTATS2025 day 3 keynote by Akshay Krishnamurthy about how to do theory research on inference time compute 👍 @aistats_conf
2
9
67
#AISTATS2025 day 3 keynote by Akshay Krishnamurthy about how to do theory research on inference time compute 👍 @aistats_conf
1
11
139
Our work on language model self-improvement will appear as an Oral at ICLR! See you in Singapore! https://t.co/eE7akC67wK
Given a high-quality verifier, language model accuracy can be improved by scaling inference-time compute (e.g., w/ repeated sampling). When can we expect similar gains without an external verifier? New paper: Self-Improvement in Language Models: The Sharpening Mechanism
1
12
94
Given a high-quality verifier, language model accuracy can be improved by scaling inference-time compute (e.g., w/ repeated sampling). When can we expect similar gains without an external verifier? New paper: Self-Improvement in Language Models: The Sharpening Mechanism
3
48
254
Check out the paper for more details: https://t.co/BHAp05pxen Joint work w/ Audrey Huang (@auddery), Adam Block, Dhruv Rohatgi, Cyril Zhang (@_cyrilzhang), Max Simchowitz (@max_simchowitz), Jordan Ash (@jordan_t_ash), and Akshay Krishnamurthy
arxiv.org
Recent work in language modeling has raised the possibility of self-improvement, where a language models evaluates and refines its own generations to achieve higher performance without external...
1
2
18
New preprint: Is Behavior Cloning All You Need? Understanding Horizon in Imitation Learning We show that good old fashioned behavior cloning enjoys horizon-independent sample complexity for imitation learning—provided you use the log loss! https://t.co/Z05IFVIuYa Thread below
arxiv.org
Imitation learning (IL) aims to mimic the behavior of an expert in a sequential decision making task by learning from demonstrations, and has been widely applied to robotics, autonomous driving,...
2
28
86
New work on understanding preference fine-tuning/RLHF -- we analyze online and offline preference fine-tuning methods via the theoretical tool of dataset coverage and reveal the importance of online unlabeled data. Plus, a new algorithm! (1/n)
1
15
93