TY - JOUR
T1 - Active Multi-Fidelity Bayesian Online Changepoint Detection
AU - Gundersen, Gregory W.
AU - Cai, Diana
AU - Zhou, Chuteng
AU - Engelhardt, Barbara E.
AU - Adams, Ryan P.
N1 - Funding Information:
We thank Paul Whatmough and Igor Fedorov for helpful conversations on ML for resource-constrained devices. B.E. Engelhardt and G.W. Gundersen received support from a grant from the Helmsley Trust, a grant from the NIH HTAN Research Program, NIH NHLBI R01 HL133218, and NSF CAREER AWD1005627. D. Cai was supported in part by a Google Ph.D. Fellowship in Machine Learning. R.P. Adams was supported in part by NSF IIS-2007278.
Publisher Copyright:
© 2021 Proceedings of Machine Learning Research. All Rights Reserved.
PY - 2021
Y1 - 2021
N2 - Online algorithms for detecting changepoints, or abrupt shifts in the behavior of a time series, are often deployed with limited resources, e.g., to edge computing settings such as mobile phones or industrial sensors. In these scenarios it may be beneficial to trade the cost of collecting an environmental measurement against the quality or “fidelity” of this measurement and how the measurement affects changepoint estimation. For instance, one might decide between inertial measurements or GPS to determine changepoints for motion. A Bayesian approach to changepoint detection is particularly appealing because we can represent our posterior uncertainty about changepoints and make active, cost-sensitive decisions about data fidelity to reduce this posterior uncertainty. Moreover, the total cost could be dramatically lowered through active fidelity switching, while remaining robust to changes in data distribution. We propose a multi-fidelity approach that makes cost-sensitive decisions about which data fidelity to collect based on maximizing information gain with respect to changepoints. We evaluate this framework on synthetic, video, and audio data and show that this information-based approach results in accurate predictions while reducing total cost.
AB - Online algorithms for detecting changepoints, or abrupt shifts in the behavior of a time series, are often deployed with limited resources, e.g., to edge computing settings such as mobile phones or industrial sensors. In these scenarios it may be beneficial to trade the cost of collecting an environmental measurement against the quality or “fidelity” of this measurement and how the measurement affects changepoint estimation. For instance, one might decide between inertial measurements or GPS to determine changepoints for motion. A Bayesian approach to changepoint detection is particularly appealing because we can represent our posterior uncertainty about changepoints and make active, cost-sensitive decisions about data fidelity to reduce this posterior uncertainty. Moreover, the total cost could be dramatically lowered through active fidelity switching, while remaining robust to changes in data distribution. We propose a multi-fidelity approach that makes cost-sensitive decisions about which data fidelity to collect based on maximizing information gain with respect to changepoints. We evaluate this framework on synthetic, video, and audio data and show that this information-based approach results in accurate predictions while reducing total cost.
UR - http://www.scopus.com/inward/record.url?scp=85163308771&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85163308771&partnerID=8YFLogxK
M3 - Conference article
AN - SCOPUS:85163308771
SN - 2640-3498
VL - 161
SP - 1916
EP - 1926
JO - Proceedings of Machine Learning Research
JF - Proceedings of Machine Learning Research
T2 - 37th Conference on Uncertainty in Artificial Intelligence, UAI 2021
Y2 - 27 July 2021 through 30 July 2021
ER -