Linzner, D. ; Koeppl, H. (2021)
Active Learning of Continuous-time Bayesian Networks through Interventions.
38th International Conference on Machine Learning. virtual Conference (18.-24.07.2021)
Konferenzveröffentlichung, Bibliographie
Dies ist die neueste Version dieses Eintrags.
Kurzbeschreibung (Abstract)
We consider the problem of learning structures and parameters of Continuous-time Bayesian Networks (CTBNs) from time-course data under minimal experimental resources. In practice, the cost of generating experimental data poses a bottle-neck, especially in the natural and social sciences. A popular approach to overcome this is Bayesian optimal experimental design (BOED). However, BOED becomes infeasible in high-dimensional settings, as it involves integration over all possible experimental outcomes. We propose a novel criterion for experimental design based on a variational approximation of the expected information gain. We show that for CTBNs, a semi-analytical expression for this criterion can be calculated for structure and parameter learning. By doing so, we can replace sampling over experimental outcomesby solving the CTBNs master-equation, for which scalable approximations exist. This all eviates the computational burden of integrating over possible experimental outcomes in high-dimensions. We employ this framework in order to recommend interventional sequences. In this context, we extend the CTBN model to conditional CTBNs in order to incorporate interventions. We demonstrate the performance of our criterion on synthetic and real-world data.
Typ des Eintrags: | Konferenzveröffentlichung |
---|---|
Erschienen: | 2021 |
Autor(en): | Linzner, D. ; Koeppl, H. |
Art des Eintrags: | Bibliographie |
Titel: | Active Learning of Continuous-time Bayesian Networks through Interventions |
Sprache: | Englisch |
Publikationsjahr: | 19 Juli 2021 |
Verlag: | ML Research Press |
Buchtitel: | Proceedings of the 38th International Conference on Machine Learning |
Veranstaltungstitel: | 38th International Conference on Machine Learning |
Veranstaltungsort: | virtual Conference |
Veranstaltungsdatum: | 18.-24.07.2021 |
URL / URN: | https://icml.cc/Conferences/2021 |
Zugehörige Links: | |
Kurzbeschreibung (Abstract): | We consider the problem of learning structures and parameters of Continuous-time Bayesian Networks (CTBNs) from time-course data under minimal experimental resources. In practice, the cost of generating experimental data poses a bottle-neck, especially in the natural and social sciences. A popular approach to overcome this is Bayesian optimal experimental design (BOED). However, BOED becomes infeasible in high-dimensional settings, as it involves integration over all possible experimental outcomes. We propose a novel criterion for experimental design based on a variational approximation of the expected information gain. We show that for CTBNs, a semi-analytical expression for this criterion can be calculated for structure and parameter learning. By doing so, we can replace sampling over experimental outcomesby solving the CTBNs master-equation, for which scalable approximations exist. This all eviates the computational burden of integrating over possible experimental outcomes in high-dimensions. We employ this framework in order to recommend interventional sequences. In this context, we extend the CTBN model to conditional CTBNs in order to incorporate interventions. We demonstrate the performance of our criterion on synthetic and real-world data. |
Freie Schlagworte: | Machine Learning |
Zusätzliche Informationen: | Erstveröffentlichung |
Fachbereich(e)/-gebiet(e): | 18 Fachbereich Elektrotechnik und Informationstechnik 18 Fachbereich Elektrotechnik und Informationstechnik > Institut für Nachrichtentechnik > Bioinspirierte Kommunikationssysteme 18 Fachbereich Elektrotechnik und Informationstechnik > Institut für Nachrichtentechnik |
Hinterlegungsdatum: | 22 Jun 2021 06:02 |
Letzte Änderung: | 03 Jul 2024 02:52 |
PPN: | |
Export: | |
Suche nach Titel in: | TUfind oder in Google |
Verfügbare Versionen dieses Eintrags
-
Active Learning of Continuous-time Bayesian Networks through Interventions. (deposited 31 Mär 2023 08:26)
- Active Learning of Continuous-time Bayesian Networks through Interventions. (deposited 22 Jun 2021 06:02) [Gegenwärtig angezeigt]
![]() |
Frage zum Eintrag |
Optionen (nur für Redakteure)
![]() |
Redaktionelle Details anzeigen |