Yan-Chao Hao

Seeing is Believing: Grounding Long-Video Understanding in Spatio-Temporal Visual Evidence

DraftMind: Efficient Video Understanding in Vision-Language Models via Chain of Draft

Large vision-language models (LVLMs) such as Qwen-VL series have demonstrated strong capabilities in visual understanding and reasoning, particularly for static images. However, extending them to video reasoning remains computationally expensive, often incurring substantial latency and token usage under conventional Ch...

CALR: Continuous Anchored Latent Reasoning via Render-of-Thought Compression

Visual latent reasoning compresses rendered derivations into compact intermediate states, reducing textual reasoning overhead. Existing approaches differ in how they represent these states: continuous methods avoid vocabulary constraints, whereas discrete methods improve accuracy through quantization into a finite code...

DraftMind: Efficient Video Understanding in Vision-Language Models via Chain of Draft

Large vision-language models (LVLMs) such as Qwen-VL series have demonstrated strong capabilities in visual understanding and reasoning, particularly for static images. However, extending them to video reasoning remains computationally expensive, often incurring substantial latency and token usage under conventional Ch...

Routing Before Looking: Query-Adaptive Evidence Acquisition for Long-form Video Understanding

Long-form video understanding remains challenging for video agents due to the mismatch between query demands and evidence acquisition strategies. Although recent planning-before-perception methods outperform query-agnostic pipelines, they often rely on a single dominant strategy, either generation-based strategy or ret...

Video-HolmesV2: Can MLLMs Reason with Spatio-Temporal Audio-Visual Evidence in Long Videos?

Video-HolmesV2 is presented, a novel benchmark designed for Deep Audio-Visual Coupling that enforces an Evidence-Based Evaluation, requiring models to justify answers with precise spatio-temporal audio-visual evidence, thereby reducing confounding effects of guessing and hallucinated evidence.

Thinking in Video: Can Video Generators Really Reason About the Real World?

The Causal-Generative Dual-Judge (CGDJ) is introduced, auditing World Model Consistency from two perspectives, and analysis exposes audio-visual misalignment, where models verbalize correct causal logic more reliably than they render it, challenging the"world simulator"narrative.

LUT: Latent Utility Training for Visual Reasoning

Experiments on perception-intensive visual reasoning benchmarks show that LUT outperforms previous latent reasoning methods and remains competitive with latent-text interleaved methods with lower annotation cost.

Beyond Surface Features: Advancing Medical Vision-Language Alignment via Dynamic Evidence-Guided Preference Optimization

Dynamic Evidence-Guided Preference Optimization (DEPO) is proposed, a new framework that enables evidence-aware and adaptive preference learning for Med-LVLMs and introduces Multi-Modal Evidence Perturbation (MEP) to suppress non-causal textual and visual shortcuts and Dispre-ferred Evidence Resampling (DER) to continu...