Hi there 👋

PathOnAI.org is a 501(c)(3) open-source non-profit organization with a mission to advance open-source development and AGI research through inclusive collaboration, mentorship, and community-driven innovation. We aim to make cutting-edge artificial general intelligence research accessible to all, fostering a diverse ecosystem where knowledge is freely shared.

Join the community!

1. Publications

• VisualTreeSearch: An Intuitive Interface for Understanding Web Agent Decision Processes
  📝 Danqing Zhang, Yaoyao Qian, Shiying He, Yuanli Wang, Jingyi Ni, Junyu Cao
  📄 ECML-PKDD'25: European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML-PKDD 2025)
  🔗 Paper
• LiteWebAgent: The Open-Source Suite for VLM-Based Web-Agent Applications
  📝 Danqing Zhang, Balaji Rama, Jingyi Ni, Shiying He, Fu Zhao, Kunyu Chen, Arnold Chen, Junyu Cao
  📄 NAACL'25: 2025 Annual Conference of the North American Chapter of the Association for Computational Linguistics – System Demonstration Track
  🔗 Paper

2. Current Projects

• in preparation

3. Past Projects

• ARC AGI Experiments
  🧠 Exploring advanced reasoning capabilities in artificial intelligence systems.

• Awesome ARC AGI
  📚 Collection of resources on advanced reasoning and cognitive architectures for AGI.

• 3D Embodied Agent
  🎮 Developing embodied AI agents for Roblox games.

• Awesome 3D Embodied AI
  📚 Curated list of resources for 3D embodied AI research and development.

• LiteMultiAgent
  👥 A library for LLM-based multi-agent applications.

• LiteWebAgent
  🌍 A library for LLM-based web-agent applications.

• LiteGUIAgent
  🤖 A library for VLM-based computer control agents.

• Awesome AI Agents
  📚 Collection of materials on AI agents.

• Natural Language Terminal
  💻 Takes a natural language command and directly calls into the OS.

4. Research Philosophy

• 🔍 Make user interfaces more intuitive for understanding and interpreting AI decision-making processes.
• 🧮 Formulate real-world applications into mathematical models to find optimal solutions.
• 🔄 Bridge the gap between academic research and production applications by raising the bar on code quality.
• 📊 Conduct detailed ablation studies to examine when and why approaches work better.