23 citations · 38 across the 13 of their papers we have counts for
17 papers
RL-IF: Tracking Non-local Training Data Influence for Online Reinforcement Learning
Shixuan Liu, Cheng Tang, Yuzheng Hu +3
Online reinforcement learning (RL) has achieved significant success in decision-making tasks, but modern online RL remains highly sensitive to the quality of training experience. U…
SHIP: Structured and Hierarchical Insight with Privacy
Kai Chen, Guisi Ni, Albert Cheu +4
Extracting insights from textual data while protecting user privacy is increasingly important. Existing methods produce private insights in diverse forms for different analytical p…
MAPLE: Metadata Augmented Private Language Evolution
Eli Chien, Yuzheng Hu, Ryan McKenna +3
Differentially private (DP) fine-tuning of large language models (LLMs) requires massive compute and full model access, which rules out state-of-the-art proprietary APIs for genera…
A Unified Theory of Random Projection for Influence Functions
Pingbang Hu, Yuzheng Hu, Jiaqi W. Ma +1
Influence functions and related data attribution scores take the form of , where is a curvature operator. In modern overparametrized models,…
OPUS: Towards Efficient and Principled Data Selection in Large Language Model Pre-training in Every Iteration
Shaobo Wang, Xuan Ouyang, Tianyi Xu +9
As high-quality public text approaches exhaustion, a phenomenon known as the Data Wall, pre-training is shifting from more tokens to better tokens. However, existing methods either…
ACTG-ARL: Differentially Private Conditional Text Generation with RL-Boosted Control
Yuzheng Hu, Ryan McKenna, Da Yu +4
Generating high-quality synthetic text under differential privacy (DP) is critical for training and evaluating language models without compromising user privacy. Prior work on synt…
Taming Hyperparameter Sensitivity in Data Attribution: Practical Selection Without Costly Retraining
Weiyi Wang, Junwei Deng, Yuzheng Hu +5
Data attribution methods, which quantify the influence of individual training data points on a machine learning model, have gained increasing popularity in data-centric application…
A Snapshot of Influence: A Local Data Attribution Framework for Online Reinforcement Learning
Yuzheng Hu, Fan Wu, Haotian Ye +5
Online reinforcement learning (RL) excels in complex, safety-critical domains but suffers from sample inefficiency, training instability, and limited interpretability. Data attribu…
MergeBench: A Benchmark for Merging Domain-Specialized LLMs
Yifei He, Siqi Zeng, Yuzheng Hu +3
Model merging provides a scalable alternative to multi-task training by combining specialized finetuned models through parameter arithmetic, enabling efficient deployment without t…
Empirical Privacy Variance
Yuzheng Hu, Fan Wu, Ruicheng Xian +5
We propose the notion of empirical privacy variance and study it in the context of differentially private fine-tuning of language models. Specifically, we show that models calibrat…
Humanity's Last Exam
Long Phan, Alice Gatti, Ziwen Han +1168
Benchmarks are important tools for tracking the rapid advancements in large language model (LLM) capabilities. However, benchmarks are not keeping pace in difficulty: LLMs now achi…
Most Influential Subset Selection: Challenges, Promises, and Beyond
Yuzheng Hu, Pingbang Hu, Han Zhao +1
How can we attribute the behaviors of machine learning models to their training data? While the classic influence function sheds light on the impact of individual samples, it often…
Localize-and-Stitch: Efficient Model Merging via Sparse Task Arithmetic
Yifei He, Yuzheng Hu, Yong Lin +2
Model merging offers an effective strategy to combine the strengths of multiple finetuned models into a unified model that preserves the specialized capabilities of each. Existing…
Revisiting Scalarization in Multi-Task Learning: A Theoretical Perspective
Yuzheng Hu, Ruicheng Xian, Qilong Wu +3
Linear scalarization, i.e., combining all loss functions by a weighted sum, has been the default choice in the literature of multi-task learning (MTL) since its inception. In recen…
Understanding the Impact of Adversarial Robustness on Accuracy Disparity
Yuzheng Hu, Fan Wu, Hongyang Zhang +1
While it has long been empirically observed that adversarial robustness may be at odds with standard accuracy and may have further disparate impacts on different classes, it remain…
Actor-critic is implicitly biased towards high entropy optimal policies
Yuzheng Hu, Ziwei Ji, Matus Telgarsky
We show that the simplest actor-critic method -- a linear softmax policy updated with TD through interaction with a linear MDP, but featuring no explicit regularization or explorat…
Towards Understanding the Data Dependency of Mixup-style Training
Muthu Chidambaram, Xiang Wang, Yuzheng Hu +2
In the Mixup training paradigm, a model is trained using convex combinations of data points and their associated labels. Despite seeing very few true data points during training, m…