Machine Learning has proven its immense potential in analyzing high dimensional data as in pictures, in digesting huge quantities of data as in videos and in controlling highly complex systems as seen in games of Go. However, in many ways leveraging the technology"s potential to create tangible value in the industrial setting is challenging. Use cases of interest in the industrial setting often vary strongly from use cases in the academic field and the requirements on the solutions are often harsher in data efficiency in reliability. I will present how Bosch is combining existing domain and engineering knowledge with ML to meet this challenge in order to become a truly AI-powered company.
Dr. Christian Daniel received his PhD from TU Darmstadt in 2016 working with Jan Peters on reinforcement learning algorithms for dynamic robot skills and has worked at Microsoft Research in Cambridge before joining Bosch. Since 2019, Christian Daniel is Head of Reinforcement Learning and Planning at the Bosch Center for AI (BCAI), focusing on bridging the gap between academic excellence and impact in industrial applications. Since 2021 Christian Daniel has taken on the role of Strategic Portfolio Coordinator for AI at BCAI, further strengthening Bosch’s focus on industrializing the potential of AI technology for the connected products of the future.
Probabilistic Movement Primitives by A Paraschos, C Daniel, J Peters, G Neumann in Advances in Neural Information Processing Systems, 2013
Hierarchical Relative Entropy Policy Search by C Daniel, G Neumann, J Peters in Artificial Intelligence and Statistics, 2012
Towards Learning Hierarchical Skills for Multi-phase Manipulation Tasks by O Kroemer, C Daniel, G Neumann, H Van Hoof, J Peters in IEEE International Conference on Robotics and Automation (ICRA), 2015