Sirui Han

EgoTwin: Dreaming Body and View in First Person

Motion-R1: Enhancing Motion Generation with Decomposed Chain-of-Thought and RL Binding

WristWorld: Generating Wrist-Views via 4D World Models for Robotic Manipulation

EffiVMT: Video Motion Transfer via Efficient Spatial-Temporal Decoupled Finetuning

CommonKV: Compressing KV Cache with Cross-layer Parameter Sharing

More Bang for the Buck: Process Reward Modeling with Entropy-Driven Uncertainty

MedInsightBench: Evaluating Medical Analytics Agents Through Multi-Step Insight Discovery in Multimodal Medical Data

Optimizing Temporal and Spatial Efficiency for Chain-of-Thought Reasoning in Large Language Models

Drop or Merge? Hybrid MoE LLMs Compressors via Metric-Driven Adaptive Allocation

WoW: Scaling Embodied Omni-World Model For Generalizable Manipulation Simulation

Learning Utility‑Calibrated Routing for Hierarchical Multi-Agents in Portfolio Decision‑Making

BTC-LLM: Efficient Sub-1-Bit LLM Quantization via Learnable Transformation and Binary Codebook

GSPR: Aligning LLM Safeguards as Generalizable Safety Policy Reasoners

Bit-by-Bit: Progressive QAT with Outlier Channel Splitting for Stable Low-Bit LLMs

Mitigating Deceptive Alignment via Self-Monitoring

Can World Models Benefit VLMs for World Dynamics?

Generative RLHF-V: Learning Principles from Multi-modal Human Preference

Safe RLHF-V: Safe Reinforcement Learning from Multi-modal Human Feedback

Semantic-guided Diverse Decoding for Large Language Model