Paulo E. Rauber
I am particularly interested in studying four major challenges faced by reinforcement learning agents:
- Learning state representations that enable good decisions in partially observable environments;
- Building an environment model that allows evaluating the consequences of action sequences;
- Exploring alternative ways to achieve goals considering environment model uncertainty;
- Generalizing experience across environments and bootstrapping from a curriculum.
My earlier work was mostly on the intersection between image analysis, machine learning, and information visualization.
My curriculum vitae is available here, and some of my work is available here.
If you want to talk about my academic interests, send an e-mail to pr at this domain.