Machine Learning, Robotics
TU Darmstadt, FG IAS,
Hochschulstr. 10, 64289 Darmstadt
Room E304, Building S2|02
Boris Belousov is interested in the theory and applications of motor skill learning in robotics. Insights from classical control, information theory, and statistics are bringing us to the next level of autonomy, enabling robots to perform increasingly delicate, skilled tasks. However, our understanding of generalization and compositionality in movement generation and execution is lagging behind. If you want to bridge the gap, get in touch.
Boris Belousov joined IAS in February 2016 as a member of the SKILLS4ROBOTS team committed to endowing robots with a higher level of intelligence. Prior to that, he received a MSc degree in communications and multimedia engineering from FAU and a BSc degree in applied mathematics and physics with focus on electrical engineering and cybernetics from MIPT.
Belousov, B.; Peters, J. (submitted). f-Divergence constrained policy improvement, Submitted to the Journal of Machine Learning Research (JMLR).
See Details [Details] Download Article [PDF] BibTeX Reference [BibTex]
Belousov, B.; Neumann, G.; Rothkopf, C.; Peters, J. (2016). Catching heuristics are optimal control policies, Advances in Neural Information Processing Systems (NIPS). See Details [Details] Download Article [PDF] BibTeX Reference [BibTex]