8 papers
More Reasoning Is Not Always Better: Executing Clinical Guidelines with LLMs and Rules
Xiaoqian Lu, Guangfu Guo, Long Cheng +1
Large language models (LLMs) have shown considerable promise in clinical decision support. However, existing approaches often rely on multi-step reasoning or multi-agent collaborat…
Visual Representation and History Modeling for Navigation World Models
Guangfu Guo, Xiaoqian Lu, Rui Liu +3
Navigation World Models (NWMs) predict action-conditioned visual futures for planning. Two practical challenges are central to their design: selecting a suitable visual representat…
DualOPSD: Closing the Feedback Loop in On-Policy Self-Distillation
Yutong Chen, Guangfu Guo, Zhichao Xu +1
On-policy self-distillation (OPSD) combines student generated trajectories with dense supervision from a privileged teacher of the same model. Its pointwise clipping protects the s…
DeltaFlow: Noise-Adaptive Bidirectional Gated Delta Networks for Embedded Language Flows
Guangfu Guo, Xiaoqian Lu, Linsey Pang +4
Embedded Language Flows (ELF) rely primarily on full non-causal attention for iterative denoising, repeatedly incurring quadratic sequence-mixing cost at each sampling step. Gated…
SSR: Similarity-Shift Refinement for Training-Free Object-Centric Masks
Xiaoqian Lu, Guangfu Guo
Object-centric models often produce fragmented masks, boundary leakage, and incorrect region merging. We introduce Similarity-Shift Refinement (SSR), a training-free post-hoc metho…
Beyond Static Visual Tokens: Structured Sequential Visual Chain-of-Thought Reasoning
Guangfu Guo, Xiaoqian Lu, Yue Feng +1
Current multimodal LLMs encode images as static visual prefixes and rely on text-based reasoning, lacking goal-driven and adaptive visual access. Inspired by human visual perceptio…
Med-VRAgent: A Framework for Medical Visual Reasoning-Enhanced Agents
Guangfu Guo, Xiaoqian Lu, Yue Feng
Visual Language Models (VLMs) achieve promising results in medical reasoning but struggle with hallucinations, vague descriptions, inconsistent logic and poor localization. To addr…
Structured Outputs Enable General-Purpose LLMs to be Medical Experts
Guangfu Guo, Kai Zhang, Bryan Hoo +4
Medical question-answering (QA) is a critical task for evaluating how effectively large language models (LLMs) encode clinical knowledge and assessing their potential applications…