Active Deep Probabilistic Subsampling

Onderzoeksoutput: Hoofdstuk in Boek/Rapport/CongresprocedureConferentiebijdrageAcademicpeer review

12 Citaten (Scopus)
111 Downloads (Pure)

Samenvatting

Subsampling a signal of interest can reduce costly data transfer, battery drain, radiation exposure and acquisition time in a wide range of problems. The recently proposed Deep Probabilistic Subsampling (DPS) method effectively integrates subsampling in an end-to-end deep learning model, but learns a static pattern for all datapoints. We generalize DPS to a sequential method that actively picks the next sample based on the information acquired so far; dubbed Active-DPS (A-DPS). We validate that A-DPS improves over DPS for MNIST classification at high subsampling rates. Moreover, we demonstrate strong performance in active acquisition Magnetic Resonance Image (MRI) reconstruction, outperforming DPS and other deep learning methods.
Originele taal-2Engels
Titel38th International Conference on Machine Learning
RedacteurenMarina Meila, Tong Zhang
UitgeverijPMLR
Pagina's10509-10518
Aantal pagina's10
ISBN van elektronische versie9781713845065
StatusGepubliceerd - jul. 2021
Evenement38th International Conference on Machine Learning (ICML 2021) - Virtual
Duur: 18 jul. 202124 jul. 2021
Congresnummer: 38

Publicatie series

NaamProceedings of Machine Learning Research
Volume139
ISSN van elektronische versie2640-3498

Congres

Congres38th International Conference on Machine Learning (ICML 2021)
Verkorte titelICML 2021
Periode18/07/2124/07/21

Vingerafdruk

Duik in de onderzoeksthema's van 'Active Deep Probabilistic Subsampling'. Samen vormen ze een unieke vingerafdruk.

Citeer dit