Welcome to the Executable Language Grounding (XLANG) Lab! We are part of the HKU NLP Group at the University of Hong Kong. XLang focuses on building language model agents that transform (“grounding”) language instructions into code or actions executable in real-world environments, including databases (data agent), web applications (plugins/web agent), and the physical world (robotic agent) etc,. It lies at the heart of language model agents or natural language interfaces that can interact with and learn from these real-world environments to facilitate human interaction with data analysis, web applications, and robotic instruction through conversation. Recent advances in XLang incorporate techniques such as LLM + external tools, code generation, semantic parsing, and dialog or interactive systems.
XLANG Lab
Pinned Loading
Repositories
- Qwen-CUA Public
Qwen-CUA: Native Computer Use for (Almost) Everything — a screenshot-driven agent that operates computers with keyboard and mouse, jointly developed by the Qwen Team and XLang Lab.
- OSWorld Public
[NeurIPS 2024] OSWorld: Benchmarking Multimodal Agents for Open-Ended Tasks in Real Computer Environments
- CUA-Gym Public
Scalable pipeline for synthesizing verifiable RLVR training data for computer-use agents
- CUA-Gym-Hub Public
CUA-Gym-Hub: mock web apps as reproducible RL training environments for computer-use agents
- FineVLA Public
Scalable annotation pipeline for action-aglined fine-grained instruciton for Visual-language-Action model
- osworld_image Public
- OSWorld-G Public
[NeurIPS 2025 Spotlight] Scaling Computer-Use Grounding via UI Decomposition and Synthesis
Top languages
Loading…
Most used topics
Loading…