###Connected AI and Continuous Lifelong Learning for Real-World Robots
Project ID: 2228bd1199 (You will need this ID for your application)
Research Theme: Artificial Intelligence and Robotics
UCL Lead department: Computer Science
Lead Supervisor: Zhibin Li
Project Summary:
Research Vision To develop novel distributed AI algorithms that learn cross-platform, shared robotic motor skills from rich multi-sensory data (proprioceptive memory including electronic skins), building on top of the existing massive computer vision database. This Robotics & AI research aims to exploit parallel computing in the cloud to process massive rich sensing data, and learn world models and robot motor skills that can be shared and adapted to various robots globally through wireless networks.
Research objectives (1) Develop rich soft-embodied multi-sensing for robots which is beyond the state of the art as multimodal feedback, and explore information extraction and representation through memory; (2) Learn contextual understanding of the environment, leveraging distributed vision systems; (3) Building on (1) and (2), develop novel learning algorithms to learn generic world models and robots’ physical interaction skills; (4) Develop distributed computing to cross-utilize collected data, share learned models, and adapt motor skills to a new robot body effectively using small real-world data.
Impact & significance The above capabilities in the objectives are fundamental building blocks to create an entry-point for real-world deployment of Robotics and AI Technologies – the catalyst to create unprecedented driving power and translate research into real productivity, potentially creating UCL spinouts.
Joining UCL recently this year, I envision creating this new direction as the next step of my career aspiration. I bring my expertise in real-world Robotics and AI, and work together with colleagues from MSD group and the department. I will use this opportunity to investigate this new field, create a paradigm shift and generate a scalable infrastructure for lifelong learning on real robots, which brings disruptive boosts in productivity for the future economy. This studentship serves as a springboard for me to start pioneer work towards grant proposals, and prepare for UKRI Program Grant and EU grant applications.