Název: DEnFi: Deep Ensemble Filter for Active Learning
Autoři: Ulrych, Lukáš ; Šmídl, Václav
Typ dokumentu: Výzkumné zprávy
Rok: 2020
Jazyk: eng
Edice: Research Report, svazek: 2383
Abstrakt: Deep Ensembles proved to be a one of the most accurate representation of uncertainty for deep neural networks. Their accuracy is beneficial in the task of active learning where unknown samples are selected for labeling based on the uncertainty of their prediction. Underestimation of the predictive uncertainty leads to poor exploration of the method. The main issue of deep ensembles is their computational cost since multiple complex networks have to be computed in parallel. In this paper, we propose to address this issue by taking advantage of the recursive nature of active learning. Specifically, we propose several methods how to generate initial values of an ensemble based of the previous ensemble. We provide comparison of the proposed strategies with existing methods on benchmark problems from Bayesian optimization and active classification. Practical benefits of the approach is demonstrated on example of learning ID of an IoT device from structured data using deep-set based networks.
Klíčová slova: Deep Ensembles; neural networks; uncertainty
Číslo projektu: GA18-21409S (CEP)
Poskytovatel projektu: GA ČR

Instituce: Ústav teorie informace a automatizace AV ČR (web)
Informace o dostupnosti dokumentu: Dokument je dostupný na externích webových stránkách.
Externí umístění souboru: http://library.utia.cas.cz/separaty/2020/AS/smidl-0531426.pdf
Původní záznam: http://hdl.handle.net/11104/0310095

Trvalý odkaz NUŠL: http://www.nusl.cz/ntk/nusl-438262


Záznam je zařazen do těchto sbírek:
Věda a výzkum > AV ČR > Ústav teorie informace a automatizace
Zprávy > Výzkumné zprávy