Under the supervision of Prof. Osmar Zaïane
Department of Computing Science, University of Alberta 🇨🇦

We are a research group in Machine Learning and Robotics, exploring how intelligent systems can learn, plan, and adapt through language, perception, and interaction.
Our mission is to build embodied AI agents capable of reasoning and acting in dynamic environments through multimodal understanding and reinforcement learning.

• LLM-driven Reward Design — using large language models to shape and explain RL rewards, enabling systems to adapt to faults and dynamic environmental changes
• Multi-Agent LLM Systems — enabling cooperation, communication, and planning among autonomous agents
• Language-Conditioned Robot Policies — grounding robot control in natural language instructions
• Spatial Intelligence in VLMs/LLMs — improving spatial reasoning and perception for embodied tasks
• LLM-based Planning & Skill Composition — integrating symbolic and learned reasoning through skill libraries

Combining LLMs/VLMs/VLAs with RL and Skill Libraries to advance adaptive, generalizable robotic intelligence.

Our work bridges:

• Perception → understanding visual and spatial cues
• Language → interpreting and reasoning with instructions
• Action → executing precise robotic control and planning

👉 Repositories
👉 Prof. Osmar Zaïane’s Page

Teaching machines not just to think — but to understand, imagine, and act.