Speaker: Jing Dong Title: Privacy preserving in reinforcement learning

I will first talk about the definition of privacy in reinforcement learning. Then I will introduce some recent results on upper and lower bounds while satisfying joint different privacy or local differential privacy.

Reference: https://arxiv.org/pdf/2009.09052.pdf

https://arxiv.org/pdf/2010.07778.pdf

https://arxiv.org/pdf/1901.10634.pdf

https://arxiv.org/pdf/2110.10133.pdf