Research observatory
A research encyclopedia for generative vision, video understanding, and sports AI
Build traceable reading paths from representative papers, method families, and evidence audits.
How to read
Every judgment returns to source notes with preserved evidence and audit status.
Start with the Sports AI route →Active topics19 entriesContinuously updated
Evidence coverage266 source notesFull-text and analysis status
Latest synthesisClaims / QuestionsOpen current judgments
Research route
From foundational representations to deployable systems
Recently updated
Continue through the research questions
- Index2026-07-16
- Image-Text Retrieval Review Protocol2026-07-16
- Image-Text Retrieval 40-Paper Evidence Matrix2026-07-16
- Image-Text Retrieval2026-07-16
- Video Understanding2026-07-16
- MIEB：大规模图像 embedding 评测基准（Massive Image Embedding Benchmark）2026-07-16
- 多语言多样性改善视觉语言表示（Multilingual Diversity Improves Vision-Language Representations）2026-07-16
- No Zero-Shot Without Exponential Data：预训练概念频率决定多模态模型性能2026-07-16