research themes

autonomous robots for environmental monitoring

robots for scientific exploration

visual exploration

visual similarity search

contrastive representation learning

3d vision and perception

active sampling and information gathering

collaborative human-robot environmental monitoring

field robotics

machine learning for planning and control

vision-based imitation learning

bayesian inverse reinforcement learning

planning and control under uncertainty in cost and dynamics

shared human-robot autonomy

continual/lifelong learning for control

differentiable simulators

simulation-based inference

system identification

safe robot learning

safe exploration in reinforcement learning

physical adversarial examples in robotics

backwards reachability

robust control

safe system identification

news

Feb 2021: Three papers accepted at ICRA, one on continual model-based RL, one on reachability based exploration in RL, and one on handling imperfect demonstrations in imitation and RL.
Feb 2021: One paper accepted at CVPR, on latent space disentanglement in GANs for image synthesis.
Jan 2021: Three papers accepted at ICLR, one on safe reinforcement learning, one on learning transferable skills for hierarchical planning, and one on differentiable physics and rendering simulators.
Dec 2020: Florian is co-organizing the NeurIPS 2020 workshop on differentiable computer vision, graphics, and physics.
Dec 2020: Our paper on encouraging diversity in neural network ensembles was accepted at AAAI.
Jul 2020: Our work, led by collaborators Travis, Juan Camilo, Stefan and others, won the Best Paper Award in the RSS’20 Workshop on Self-Supervised Robot Learning.
Jul 2020: Two papers accepted at IROS. One on visual search and one on mobile manipulation.
Jun 2020: Florian co-organized the Debates on the Future of Robotics Research workshop at ICRA 2020.
May 2020: Our paper on vision-based navigation was accepted at RSS.
Apr 2020: Our paper on combining gradient-based optimization with CEM was accepted at L4DC.