Repository logo
 

Latent Generative Replay for Resource-Efficient Continual Learning of Facial Expressions

Simple item page
cam.depositDate2022-10-10
dc.contributor.authorStoychev, Samuil
dc.contributor.authorChuramani, Nikhil
dc.contributor.authorGunes, Hatice
dc.contributor.orcidChuramani, Nikhil [0000-0001-5926-0091]
dc.date.accessioned2022-10-14T23:30:24Z
dc.date.available2022-10-14T23:30:24Z
dc.date.updated2022-10-10T12:42:55Z
dc.description.abstractReal-world Facial Expression Recognition (FER) systems require models to constantly learn and adapt with novel data. Traditional Machine Learning (ML) approaches struggle to adapt to such dynamics as models need to be re-trained from scratch with a combination of both old and new data. Replay-based Continual Learning (CL) provides a solution to this problem, either by storing previously seen data samples in memory, sampling and interleaving them with novel data (rehearsal) or by using a generative model to simulate pseudo- samples to replay past knowledge (pseudo-rehearsal). Yet, the high memory footprint of rehearsal and the high computational cost of pseudo-rehearsal limit the real-world application of such methods, especially on resource-constrained devices. To address this, we propose Latent Generative Replay (LGR) for pseudo-rehearsal of low-dimensional latent features to mitigate forgetting in a resource-efficient manner. We adapt popular CL strategies to use LGR instead of generating pseudo-samples, resulting in performance upgrades when evaluated on the CK+, RAF-DB and AffectNet FER benchmarks where LGR significantly reduces the memory and resource consumption of replay-based CL without compromising model performance.
dc.description.sponsorshipEPSRC
dc.identifier.doi10.17863/CAM.89541
dc.identifier.urihttps://www.repository.cam.ac.uk/handle/1810/342122
dc.language.isoeng
dc.publisher.departmentDepartment of Computer Science and Technology
dc.rightsAttribution 4.0 International
dc.rights.urihttps://creativecommons.org/licenses/by/4.0/
dc.titleLatent Generative Replay for Resource-Efficient Continual Learning of Facial Expressions
dc.typeConference Object
dcterms.dateAccepted2022-09-11
pubs.conference-finish-date2023-01-08
pubs.conference-nameIEEE International Conference on Automatic Face and Gesture Recognition, 2023
pubs.conference-start-date2023-01-05
pubs.funder-project-idEPSRC (2107412)
pubs.funder-project-idEngineering and Physical Sciences Research Council (EP/R030782/1)
pubs.licence-display-nameApollo Repository Deposit Licence Agreement
pubs.licence-identifierapollo-deposit-licence-2-1
rioxxterms.versionAM

Files

Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
LGR_FG2023_openaccess.pdf
Size:
1.35 MB
Format:
Adobe Portable Document Format
Description:
Accepted version
Licence
https://creativecommons.org/licenses/by/4.0/
Download

Collections

Cambridge University Research Outputs