Supervision Guidance

Thanks for you interest in collaborating with me. To make it happen smoothly, please send me an email (1) with the subject e.g. [Supervision Request], describing the direction you want to work on (please refer to the next section for possible directions). (2) Please also attach your CV or webpage, and (3) cc 'yongchen at gatech dot edu' in the email. Currently, I am interested in the following directions:

• Generative Diffusion Models: interesting applications or new algorithms of diffusion models.
• Robustness of Models: design robust models inspired by diffusion models, adversarial attacks.
• AI Safety for Geneartive Models: privacy / property protection for generative models.
• Generalized Embodied AI: inject web knowledge in e.g. LLM/VLM into robot policies.
• Safe Embodied AI: explore attacks/defenses in embodiede AI.