Denis Blessing
- Variational Inference, (Hierarchical) Latent Variable Models, Multimodal Density Estimation, Mixture Models, Diffusion Models
- PhD Student
- Room: 425
- denis blessing ∂ kit edu
CS 50.19 InformatiKOM I
Adenauerring 12
76131 Karlsruhe
About me
I joined the Autonomous Learning Robot group in January 2023 as a PhD student. Before that, I received my Bachelors degree in electrical engineering and my Masters degree in computer science from KIT.My research mainly focuses on developing expressive models for variational inference (VI). VI is ubiquitous in machine learning with applications in maximum entropy reinforcement learning, Bayesian inference, meta-learning and many more.
Existing VI models often lack expressivity, e.g., are not able to represent multimodal distributions or perform poorly when used for amortized VI. More expressive models are typically not scalable to high dimensional data or lack robustness. I am particularly interested in developing methods that counteract the aforementioned problems.