9 papers
Test-Time Scaling for Diffusion Language Models with Frequency-Aware Remasking
Bowen Zuo, Yue Yu, Dongruo Zhou +1
Diffusion language models (DLMs) generate text through iterative denoising, where decisions about which tokens to commit or remask strongly affect output quality. Existing certaint…
Exploring More, Reasoning Better: Stepwise Risk-Sensitive GRPO for Diffusion Language Models
Yue Yu, Yue YU, Bowen Zuo +3
Diffusion large language models (dLLMs) generate text by denoising a sequence or successive blocks, allowing several tokens to be revealed in parallel. Reinforcement learning with…
Closing the Reflection Gap: A Free Calibration Bonus for Agentic RL
Yinglun Zhu
LLMs are increasingly deployed as agents that interact with external environments and observe feedback such as execution results, error messages, and tool outputs. A well-functioni…
Active Testing of Large Language Models via Approximate Neyman Allocation
Zeli Liu, Jiancheng Zhang, Cong Liu +1
Large language models (LLMs) require reliable evaluation from pre-training to test-time scaling, making evaluation a recurring rather than one-off cost. As model scales grow and ta…
Adaptive Test-Time Compute Allocation with Evolving In-Context Demonstrations
Bowen Zuo, Dongruo Zhou, Yinglun Zhu
While scaling test-time compute can substantially improve model performance, existing approaches either rely on static compute allocation or sample from fixed generation distributi…
Test-Time Matching: Unlocking Compositional Reasoning in Multimodal Models
Yinglun Zhu, Jiancheng Zhang, Fuzhi Tang
Frontier AI models have achieved remarkable progress, yet recent studies suggest they struggle with compositional reasoning, often performing at or below random chance on establish…
Towards Multimodal Active Learning: Efficient Learning with Limited Paired Data
Jiancheng Zhang, Yinglun Zhu
Active learning (AL) is a principled strategy to reduce annotation cost in data-hungry deep learning. However, existing AL algorithms focus almost exclusively on unimodal data, ove…
LeMix: Unified Scheduling for LLM Training and Inference on Multi-GPU Systems
Yufei Li, Zexin Li, Yinglun Zhu +1
Modern deployment of large language models (LLMs) frequently involves both inference serving and continuous retraining to stay aligned with evolving data and user feedback. Common…
Mixtraining: A Better Trade-Off Between Compute and Performance
Zexin Li, Jiancheng Zhang, Yufei Li +2
Incorporating self-supervised learning (SSL) before standard supervised learning (SL) has become a widely used strategy to enhance model performance, particularly in data-limited s…