Generative models for model based reinforcement learning
||Generative models for model based reinforcement learning|
||SNIC Small Compute|
||Emilio Jorge <email@example.com>|
||Chalmers tekniska högskola|
||2021-02-22 – 2022-03-01|
Some of the will be on developing new generative deep learning models, such as generative adversarial networks (GAN) and variational autoencoders (VAE), Normalizing Flows. We aim to develop novel approaches that are capable of producing in in a way that appropriately reflects underlying uncertainty. In reinforcement learning, generative models that incorporate uncertainty can enable efficient exploration of unknown environments without strong prior assumptions.
I also do research on more classical Bayesian Reinforcement learning that is very suited for classical CPU infrastructure.