Supervisor: Dr. Wenjie Huang ([email protected])

Co-supervisor: Dr. Fangjun Li ([email protected])

• My in-progress research aims to integrate tactile information into the multimodal system for humanoid robots. The task focuses on table manipulation as well as goal-oriented action learning for robots. Specifically, the iCub model provides stereo vision and tactile signals for richer cognition. Cognitive plausibility is key to my research, as I am interested in building robotic systems that could behave somehow like humans.

Intern Role

Specific Tasks for the Intern:

1. Build a VLA (+tactile) model for the humanoid robot in either manipulation tasks or learning actions from human demonstrations.
2. Or implement an imagination mechanism into the VLA pipeline for human-robot interaction

Minimum Skills or Background:

Python programming and AI (Neural Networks) are required

Desirable skills or Background

Understanding VLAs or other foundation models is preferred but not necessary; MuJoCo experience is also a bonus.

Skills/Knowledge the Intern Will Gain at the end of the internship:

• Practical skills in foundation models, Simulator environment, robot experiment design, and hand-on experience of arm or humanoid robots.

Supervision / Mentorship Provided by Supervisor

• We would have regular meetings, usually weekly, but might have more depending on your motivation and progress. I would provide one-on-one training, providing advice on literature and the required techniques. When you get hands-on with the experiment, I would guide you through and supervise the experiment process with robots. When you arrive at the output of the experiment, I would help with the publication generation and, if we are lucky, sponsor the publication with project funding

Logistics

You are expected to work on the intern around 10 hours per week, including meetings, paper reading, coding and experiments.

Tentative Work Plan / Milestones:

1st month, learning of relevant concepts and techniques; 2nd month, experiment design; 3rd – 4th months, Model development and preliminary experiment setup; 5th – 6th months Data Collection and Model training; 7th month, results generation and analysis; 8th – 9th Month, paper writing and submission. Progress is subject to your work pace.

Expected Outcomes at the End of the Internship: