Bayesian active learning for maximal information gain on model parameters
Publikation: Bidrag til bog/antologi/rapport › Konferencebidrag i proceedings › Forskning › fagfællebedømt
Standard
Bayesian active learning for maximal information gain on model parameters. / Arnavaz, Kasra; Feragen, Aasa; Krause, Oswin; Loog, Marco.
Proceedings of ICPR 2020 - 25th International Conference on Pattern Recognition. IEEE, 2020. s. 10524-10531 9411962.Publikation: Bidrag til bog/antologi/rapport › Konferencebidrag i proceedings › Forskning › fagfællebedømt
Harvard
APA
Vancouver
Author
Bibtex
}
RIS
TY - GEN
T1 - Bayesian active learning for maximal information gain on model parameters
AU - Arnavaz, Kasra
AU - Feragen, Aasa
AU - Krause, Oswin
AU - Loog, Marco
PY - 2020
Y1 - 2020
N2 - The fact that machine learning models, despite their advancements, are still trained on randomly gathered data is proof that a lasting solution to the problem of optimal data gathering has not yet been found. In this paper, we investigate whether a Bayesian approach to the classification problem can provide assumptions under which one is guaranteed to perform at least as good as random sampling. For a logistic regression model, we show that maximal expected information gain on model parameters is a promising criterion for selecting samples, assuming that our classification model is well-matched to the data. Our derived criterion is closely related to the maximum model change. We experiment with data sets which satisfy this assumption to varying degrees to see how sensitive our performance is to the violation of our assumption in practice.
AB - The fact that machine learning models, despite their advancements, are still trained on randomly gathered data is proof that a lasting solution to the problem of optimal data gathering has not yet been found. In this paper, we investigate whether a Bayesian approach to the classification problem can provide assumptions under which one is guaranteed to perform at least as good as random sampling. For a logistic regression model, we show that maximal expected information gain on model parameters is a promising criterion for selecting samples, assuming that our classification model is well-matched to the data. Our derived criterion is closely related to the maximum model change. We experiment with data sets which satisfy this assumption to varying degrees to see how sensitive our performance is to the violation of our assumption in practice.
U2 - 10.1109/ICPR48806.2021.9411962
DO - 10.1109/ICPR48806.2021.9411962
M3 - Article in proceedings
AN - SCOPUS:85110513578
SP - 10524
EP - 10531
BT - Proceedings of ICPR 2020 - 25th International Conference on Pattern Recognition
PB - IEEE
T2 - 25th International Conference on Pattern Recognition, ICPR 2020
Y2 - 10 January 2021 through 15 January 2021
ER -
ID: 286999036