elib
DLR-Header
DLR-Logo -> http://www.dlr.de
DLR Portal Home | Impressum | Datenschutz | Kontakt | English
Schriftgröße: [-] Text [+]

Bayesian Convolutional Neural Network: Robustly Quantify Uncertainty for Misclassifications Detection

Njieutcheu Tassi, Cedrique Rovile (2019) Bayesian Convolutional Neural Network: Robustly Quantify Uncertainty for Misclassifications Detection. In: Communications in Computer and Information Science, Seiten 118-132. Springer, Cham. Third Mediterranean Conference on Pattern Recognition and Artificial Intelligence, 2019-12-22 - 2019-12-23, Türkey. doi: 10.1007/978-3-030-37548-5_10. ISBN 978-303032422-3. ISSN 1865-0929.

[img] PDF - Nur DLR-intern zugänglich
945kB

Offizielle URL: https://link.springer.com/chapter/10.1007/978-3-030-37548-5_10#citeas

Kurzfassung

For safety and mission critical systems relying on Convolutional Neural Networks (CNNs), it is crucial to avoid incorrect predictions that can cause accident or financial crisis. This can be achieved by quantifying and interpreting the predictive uncertainty. Current methods for uncertainty quantification rely on Bayesian CNNs that approximate Bayesian inference via dropout sampling. This paper investigates different dropout methods to robustly quantify the predictive uncertainty for misclassifications detection. Specifically, the following questions are addressed: In which layers should activations be sampled? Which dropout sampling mask should be used? What dropout probability should be used? How to choose the number of ensemble members? How to combine ensemble members? How to quantify the classification uncertainty? To answer these questions, experiments were conducted on three datasets using three different network architectures. Experimental results showed that the classification uncertainty is best captured by averaging the predictions of all stochastic CNNs sampled from the Bayesian CNN and by validating the predictions of the Bayesian CNN with three uncertainty measures, namely the predictive confidence, predictive entropy and standard deviation thresholds. The results showed further that the optimal dropout method specified through the sampling location, sampling mask, inference dropout probability, and number of stochastic forward passes depends on both the dataset and the designed network architecture. Notwithstanding this, I proposed to sample inputs to max pooling layers with a cascade of Multiplicative Gaussian Mask (MGM) followed by Multiplicative Bernoulli Spatial Mask (MBSM) to robustly quantify the classification uncertainty, while keeping the loss in performance low.

elib-URL des Eintrags:https://elib.dlr.de/144154/
Dokumentart:Konferenzbeitrag (Vortrag)
Titel:Bayesian Convolutional Neural Network: Robustly Quantify Uncertainty for Misclassifications Detection
Autoren:
AutorenInstitution oder E-Mail-AdresseAutoren-ORCID-iDORCID Put Code
Njieutcheu Tassi, Cedrique RovileCedrique.NjieutcheuTassi (at) dlr.deNICHT SPEZIFIZIERTNICHT SPEZIFIZIERT
Datum:18 Dezember 2019
Erschienen in:Communications in Computer and Information Science
Referierte Publikation:Ja
Open Access:Nein
Gold Open Access:Nein
In SCOPUS:Ja
In ISI Web of Science:Nein
DOI:10.1007/978-3-030-37548-5_10
Seitenbereich:Seiten 118-132
Verlag:Springer, Cham
Name der Reihe:Pattern Recognition and Artificial Intelligence
ISSN:1865-0929
ISBN:978-303032422-3
Status:veröffentlicht
Stichwörter:Convolutional Neural Networks (CNNs) Bayesian CNNs Dropout sampling Uncertainty quantification Uncertainty quality
Veranstaltungstitel:Third Mediterranean Conference on Pattern Recognition and Artificial Intelligence
Veranstaltungsort:Türkey
Veranstaltungsart:internationale Konferenz
Veranstaltungsbeginn:22 Dezember 2019
Veranstaltungsende:23 Dezember 2019
Veranstalter :The Departement of Computer Engeneering, Istanbul Sabahattin Zaim University
HGF - Forschungsbereich:Luftfahrt, Raumfahrt und Verkehr
HGF - Programm:Raumfahrt
HGF - Programmthema:Technik für Raumfahrtsysteme
DLR - Schwerpunkt:Raumfahrt
DLR - Forschungsgebiet:R SY - Technik für Raumfahrtsysteme
DLR - Teilgebiet (Projekt, Vorhaben):R - Sicherstellung der Informationsqualität in KI-Systemen (SIQKIS)
Standort: Berlin-Adlershof
Institute & Einrichtungen:Institut für Optische Sensorsysteme > Echtzeit-Datenprozessierung
Hinterlegt von: Njieutcheu Tassi, Cedrique Rovile
Hinterlegt am:30 Sep 2021 09:02
Letzte Änderung:24 Apr 2024 20:43

Nur für Mitarbeiter des Archivs: Kontrollseite des Eintrags

Blättern
Suchen
Hilfe & Kontakt
Informationen
electronic library verwendet EPrints 3.3.12
Gestaltung Webseite und Datenbank: Copyright © Deutsches Zentrum für Luft- und Raumfahrt (DLR). Alle Rechte vorbehalten.