Yuxuan Wang

Reasoning via Test-Time Instance-Level Policy Gradient in Latent Space

Omni-Captioner: Data Pipeline, Models, and Benchmark for Omni Detailed Perception

Continuous-Time Discrete Markov Bridge

OmniVideoBench: Towards Audio-Visual Understanding Evaluation for Omni MLLMs

v-HUB: A Visual-Centric Humor Understanding Benchmark for Video LLMs

Beyond Isolated Facts: Synthesizing Narrative and Grounded Supervision for VideoQA

Multi-scale Temporal Prediction via Incremental Generation and Multi-agent Collaboration

TokenSwift: Lossless Acceleration of Ultra Long Sequence Generation

LongViTU: Instruction Tuning for Long-Form Video Understanding