Hi there! π I'm Zhishan Zou (ιΉεΏε±±), a student at Beijing University of Posts and Telecommunications. My research explores artificial intelligence, with a particular focus on video understanding and generation, behavior and cognition, and agentic intelligence. You can find my publications on Google Scholar. I'm always open to research collaborations and conversations around brave new ideas.
βοΈ Self in Space Benchmarking Self-awareness and spatial cognition for UAV embodied intelligence.- π³ Poplar A Scalable Pipeline for Human-Centric Image Dataset Synthesis.
- π NBA Games Full-game video metadata with official statistics and play-by-play annotations.
- π Air Minecraft A 6-DOF voxel-world drone simulator for aerial robotics research.
βοΈ Self in Space [ACM MM 2026] Benchmarking self-awareness and spatial cognition in UAV embodied intelligence. [arXiv] [project page] [code] [huggingface] [modelscope]- π³ Poplar A Scalable Pipeline for Human-Centric Image Dataset Synthesis. [arXiv] [project page] [code] [huggingface] [modelscope]
- π OneThree arXiv Template A reusable LaTeX template for papers, technical reports, and project manuscripts. [code]
- π€ MIMO Claude Code Traces 1K A dataset of 1,017 complete Claude Code-style coding-agent trajectories generated by MiMo-V2.5-Pro. [code] [huggingface] [modelscope]
- β½ FIFA World Cup Games Full-match video metadata for 40 FIFA World Cup games with structured statistics and event timelines. [code] [huggingface] [modelscope]
- π Space Lab An Astro website template for research groups, labs, and individual researchers with an interactive 3D globe. [code]
- π Air Minecraft A 6-DOF drone simulator built on the Luanti voxel engine for aerial robotics and spatial intelligence research. [code] [video]
- π NBA Games A dataset of 189 verified full-game NBA videos linked to official box scores and play-by-play annotations. [code] [huggingface] [modelscope]
- π ICCV Papers ICCV papers from 2013 to present with metadata, abstracts, BibTeX records, and full-paper PDFs. [code] [huggingface] [modelscope]
- π Dog100K A multimodal dataset containing 103,508 high-quality dog image-text pairs. [code] [huggingface]
- π CVPR Papers CVPR papers from 2013 to present with metadata, abstracts, BibTeX records, and full-paper PDFs. [code] [huggingface] [modelscope]
- π§ Gaokao Compass 2025 An interactive admissions data platform connecting score ranks, enrollment plans, and admission results across China. [code] [huggingface]
- π CVPR 2026 Paper Explorer A bilingual visualization platform for searching, classifying, and reading structured introductions to CVPR 2026 papers. [huggingface] [modelscope]
- π Using Claude for Microsoft 365 with a Third-Party API Gateway A practical guide to connecting the Claude for Microsoft 365 add-in to a third-party LLM gateway through CLIProxyAPI.
- π LLMs Can't Jump and Abductive Reasoning An exploration of abductive reasoning, subjective experience, and why language models can extend existing explanations yet still struggle to make genuine scientific conceptual leaps.