Carlos Luis

Carlos Luis joined the Intelligent Autonomous Systems group in April 2021 as an external Ph.D. student in collaboration with the Bosch Center for Artificial Intelligence (BCAI), working on uncertainty representations for model-based reinforcement learning.

Before starting his Ph.D., he worked as a Software Development Engineer at Amazon, working on the large-scale drone delivery project known as Amazon Prime Air, where he developed safety-critical mission planning components. He holds a M.Sc. from the University of Toronto, where he did research in multi-robot motion planning at the Dynamic Systems Lab under Prof. Angela Schoellig. Please see his Curriculum Vitae for his complete trajectory.

During his Ph.D., Carlos is studying how to best represent and use uncertainty information about the environment within reinforcement learning algorithms. The goal is to improve the sample complexity of these algorithms and bring them closer to real-world applications.

Carlos has been a reviewer for IEEE Robotics and Automation Letters (IEEE RA-L)

Publications

  •     Bib
    Luis, C.E.; Bottero, A.G.; Vinogradska, J.; Berkenkamp, F.; Peters, J. (submitted). Uncertainty Representations in State-Space Layers for Deep Reinforcement Learning under Partial Observability, Transactions on Machine Learning Research (TMLR).
  •   Bib
    Luis, C.E.; Bottero, A.G.; Vinogradska, J.; Berkenkamp, F.; Peters, J. (2024). Value-Distributional Model-Based Reinforcement Learning, Journal of Machine Learning Research (JMLR).
  •     Bib
    Luis, C.; Bottero, A.G.; Vinogradska, J.; Berkenkamp, F.; Peters, J. (2023). Model-Based Uncertainty in Value Functions, Proceedings of the International Conference on Artificial Intelligence and Statistics (AISTATS).
  •     Bib
    Bottero, A.G.; Luis, C.E.; Vinogradska, J.; Berkenkamp, F.; Peters, J. (2022). Information-Theoretic Safe Exploration with Gaussian Processes, Advances in Neural Information Processing Systems (NIPS / NeurIPS).