PPaperPicks

Kun Xu

10 papers at tracked venues · 9 at CORE A* · active 20242026

Venues

Frequent coauthors

Papers

  1. Generating Attribute-Aware Human Motions from Textual Prompt
  2. Following the Autoregressive Nature of LLM Embeddings via Compression and Alignment
  3. Granularity-Adaptive Spatial Evidence Tokenization for Video Question Answering
  4. Pyramidal Flow Matching for Efficient Video Generative Modeling
  5. Harder Task Needs More Experts: Dynamic Routing in MoE Models
  6. RectifID: Personalizing Rectified Flow with Anchored Classifier Guidance
  7. Unified Language-Vision Pretraining in LLM with Dynamic Discrete Visual Tokenization
  8. Video-LaVIT: Unified Video-Language Pre-training with Decoupled Visual-Motional Tokenization