User: Guest  Login
If you experience problems opening the document, please try this link.
Original title:
Variational Bayes for Continual Learning and Time-Series Forecasting
Translated title:
Variational Bayes für kontinuierliches Lernen und Zeitreihenvorhersage
Author:
Kurle, Richard
Year:
2023
Document type:
Dissertation
Faculty/School:
TUM School of Computation, Information and Technology
Advisor:
Günnemann, Stephan (Prof. Dr.)
Referee:
Günnemann, Stephan (Prof. Dr.); Groh, Georg (Prof. Dr.)
Language:
en
Subject group:
DAT Datenverarbeitung, Informatik
TUM classification:
DAT 600; DAT 703; DAT 708
Abstract:
This thesis develops variational Bayesian methods for applications in continual learning, multi-source inference, and time-series forecasting. For continual learning, recursive approximation and adaptation methods are developed for Bayesian neural networks. Variational autoencoders are extended for multi-source learning using separate encoders and decoders for each source. For probabilistic forecasting, a Rao-Blackwellised particle filter with a variational proposal distribution is proposed.
Translated abstract:
Diese Dissertation entwickelt variational Bayesian Methoden für Anwendungen im kontinuierlichen Lernen, Inferenz aus mehreren Informationsquellen und der Zeitreihenvorhersage. Rekursive Approximations- und Adaptionsmethoden für Bayesian neural networks ermöglichen kontinuierliches Lernen. Variational autoencoders werden mit separate Encodern und Decodern für mehrere Quellen erweitert. Für Forecasting wird ein effizienter Partikel-Filter mit einer variational proposal distribution vorgeschlagen.
WWW:
https://mediatum.ub.tum.de/?id=1685412
Date of submission:
23.09.2022
Oral examination:
13.03.2023
File size:
13430724 bytes
Pages:
141
Urn (citeable URL):
https://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:bvb:91-diss-20230313-1685412-1-5
Last change:
12.04.2023
 BibTeX