|Title:||Gershgorin Loss Stabilizes the Recurrent Neural Network Compartment of an End-to-end Robot Learning Scheme||Authors:||Lechner, Mathias
|Keywords:||dynamical systems; Robot Learning; Continuous-time recurrent neural networks; Deep Learning; Machine Learning; Artificial Intelligence; neural networks||Issue Date:||31-Aug-2020||Book Title:||2020 IEEE International Conference on Robotics and Automation (ICRA)||Abstract:||
Traditional robotic control suits require profound task-specific knowledge for designing, building and testing control software. The rise of Deep Learning has enabled end-to-end solutions to be learned entirely from data, requiring minimal knowledge about the application area. We design a learning scheme to train end-to-end linear dynamical systems (LDS)s by gradient descent in imitation learning robotic domains. We introduce a new regularization loss component together with a learning algorithm that improves the stability of the learned autonomous system, by forcing the eigenvalues of the internal state updates of an LDS to be negative reals. We evaluate our approach on a series of real-life and simulated robotic experiments, in comparison to linear and nonlinear Recurrent Neural Network (RNN) architectures. Our results show that our stabilizing method significantly improves test performance of LDS, enabling such linear models to match the performance of contemporary nonlinear RNN architectures. A video of the obstacle avoidance performance of our method on a mobile robot, in unseen environments, compared to other methods can be viewed at https://youtu.be/mhEsCoNao5E.
|DOI:||10.34726/242||Organisation:||E191-01 - Forschungsbereich Cyber-Physical Systems||License:||Urheberrechtsschutz 1.0||Publication Type:||Exhibition Contribution
|Appears in Collections:||Beitrag in Konferenzband | Conference Paper|
Show full item record
Files in this item:
Items in reposiTUm are protected by copyright, with all rights reserved, unless otherwise indicated.