9 papers
Denoising as Projection: Constrained Optimization with Gradient-Guided Diffusion
Runyu Zhang, Jiawei Zhang, Gioele Zardini +2
Diffusion models are increasingly used not only for sampling from learned data distributions, but also for generating samples that optimize task-specific objectives. A common appro…
Why GRPO Needs Normalization: A Local-Curvature Perspective on Adaptive Gradients
Cheng Ge, Caitlyn Heqi Yin, Hao Liang +1
Reinforcement learning (RL) has become a key driver of language model reasoning. Among RL algorithms, Group Relative Policy Optimization (GRPO) is the de facto standard, avoiding t…
Mosaic: Unlocking Long-Context Inference for Diffusion LLMs via Global Memory Planning and Dynamic Peak Taming
Liang Zheng, Bowen Shi, Yitao Hu +5
Diffusion-based large language models (dLLMs) have emerged as a promising paradigm, utilizing simultaneous denoising to enable global planning and iterative refinement. While these…
Toward Unifying Group Fairness Evaluation from a Sparsity Perspective
Zhecheng Sheng, Jiawei Zhang, Enmao Diao
Ensuring algorithmic fairness remains a significant challenge in machine learning, particularly as models are increasingly applied across diverse domains. While numerous fairness c…
Lost in Tokenization: Context as the Key to Unlocking Biomolecular Understanding in Scientific LLMs
Kai Zhuang, Jiawei Zhang, Yumou Liu +10
Scientific Large Language Models (Sci-LLMs) have emerged as a promising frontier for accelerating biological discovery. However, these models face a fundamental challenge when proc…
Beyond RLHF and NLHF: Population-Proportional Alignment under an Axiomatic Framework
Kihyun Kim, Jiawei Zhang, Asuman Ozdaglar +1
Conventional preference learning methods often prioritize opinions held more widely when aggregating preferences from multiple evaluators. This may result in policies that are bias…
Addressing misspecification in contextual optimization
Omar Bennouna, Jiawei Zhang, Saurabh Amin +1
We study a linear contextual optimization problem where a decision maker has access to historical data and contextual features to learn a cost prediction model aimed at minimizing…
A Unified Linear Programming Framework for Offline Reward Learning from Human Demonstrations and Feedback
Kihyun Kim, Jiawei Zhang, Asuman Ozdaglar +1
Inverse Reinforcement Learning (IRL) and Reinforcement Learning from Human Feedback (RLHF) are pivotal methodologies in reward learning, which involve inferring and shaping the und…
Design-Based Causal Inference with Missing Outcomes: Missingness Mechanisms, Imputation-Assisted Randomization Tests, and Covariate Adjustment
Siyu Heng, Jiawei Zhang, Yang Feng
Design-based causal inference, also known as randomization-based or finite-population causal inference, is one of the most widely used causal inference frameworks, largely due to t…