DLR-Logo -> http://www.dlr.de
DLR Portal Home | Imprint | Privacy Policy | Contact | Deutsch
Fontsize: [-] Text [+]

Bayesian Optimization Meets Laplace Approximation for Robotic Introspection

Humt, Matthias and Lee, Jongseok and Triebel, Rudolph (2020) Bayesian Optimization Meets Laplace Approximation for Robotic Introspection. IROS 2020 Long-Term Autonomy Workshop, 25. Okt. - 25. Nov. 2020, Las Vegas, USA (online).

[img] PDF

Official URL: https://drive.google.com/file/d/1E9D8-bg-K_CIC2rlMyZg0si1zaozRfiO/view


In robotics, deep learning (DL) methods are used more and more widely, but their general inability to provide reliable confidence estimates will ultimately lead to fragile and unreliable systems. This impedes the potential deployments of DL methods for long-term autonomy. Therefore, in this paper we introduce a scalable Laplace Approximation (LA) technique to make Deep Neural Networks (DNNs) more introspective, i.e. to enable them to provide accurate assessments of their failure probability for unseen test data. In particular, we propose a novel Bayesian Optimization (BO) algorithm to mitigate their tendency of under-fitting the true weight posterior, so that both the calibration and the accuracy of the predictions can be simultaneously optimized. We demonstrate empirically that the proposed BO approach requires fewer iterations for this when compared to random search, and we show that the proposed framework can be scaled up to large datasets and architectures.

Item URL in elib:https://elib.dlr.de/137021/
Document Type:Conference or Workshop Item (Other)
Title:Bayesian Optimization Meets Laplace Approximation for Robotic Introspection
AuthorsInstitution or Email of AuthorsAuthor's ORCID iD
Humt, MatthiasMatthias.Humt (at) dlr.deUNSPECIFIED
Lee, JongseokJongseok.Lee (at) dlr.deUNSPECIFIED
Triebel, RudolphRudolph.Triebel (at) dlr.dehttps://orcid.org/0000-0002-7975-036X
Date:25 October 2020
Refereed publication:No
Open Access:Yes
Gold Open Access:No
In ISI Web of Science:No
Keywords:Bayesian Optimization, Robotics, Deep Learning, Introspection, Bayesian Neural Networks, Laplace Approximation
Event Title:IROS 2020 Long-Term Autonomy Workshop
Event Location:Las Vegas, USA (online)
Event Type:Workshop
Event Dates:25. Okt. - 25. Nov. 2020
HGF - Research field:Aeronautics, Space and Transport
HGF - Program:Space
HGF - Program Themes:Space System Technology
DLR - Research area:Raumfahrt
DLR - Program:R SY - Space System Technology
DLR - Research theme (Project):R - Vorhaben Multisensorielle Weltmodellierung (old)
Location: Oberpfaffenhofen
Institutes and Institutions:Institute of Robotics and Mechatronics (since 2013) > Perception and Cognition
Deposited By: Humt, Matthias
Deposited On:04 Nov 2020 18:15
Last Modified:04 Nov 2020 18:15

Repository Staff Only: item control page

Help & Contact
electronic library is running on EPrints 3.3.12
Copyright © 2008-2017 German Aerospace Center (DLR). All rights reserved.