Controlling Forgetting with Test-Time Data in Continual Learning
| Title: | Controlling Forgetting with Test-Time Data in Continual Learning |
|---|---|
| Authors: | Singh, Vaibhav, Aljundi, Rahaf, Belilovsky, Eugene |
| Publication Year: | 2024 |
| Collection: | Computer Science |
| Subject Terms: | Computer Science - Machine Learning |
| More Details: | Foundational vision-language models have shown impressive performance on various downstream tasks. Yet, there is still a pressing need to update these models later as new tasks or domains become available. Ongoing Continual Learning (CL) research provides techniques to overcome catastrophic forgetting of previous information when new knowledge is acquired. To date, CL techniques focus only on the supervised training sessions. This results in significant forgetting yielding inferior performance to even the prior model zero shot performance. In this work, we argue that test-time data hold great information that can be leveraged in a self supervised manner to refresh the model's memory of previous learned tasks and hence greatly reduce forgetting at no extra labelling cost. We study how unsupervised data can be employed online to improve models' performance on prior tasks upon encountering representative samples. We propose a simple yet effective student-teacher model with gradient based sparse parameters updates and show significant performance improvements and reduction in forgetting, which could alleviate the role of an offline episodic memory/experience replay buffer. Comment: 9 pages, 2 figures |
| Document Type: | Working Paper |
| Access URL: | http://arxiv.org/abs/2406.13653 |
| Accession Number: | edsarx.2406.13653 |
| Database: | arXiv |
| FullText | Text: Availability: 0 CustomLinks: – Url: http://arxiv.org/abs/2406.13653 Name: EDS - Arxiv Category: fullText Text: View this record from Arxiv MouseOverText: View this record from Arxiv – Url: https://resolver.ebsco.com/c/xy5jbn/result?sid=EBSCO:edsarx&genre=article&issn=&ISBN=&volume=&issue=&date=20240619&spage=&pages=&title=Controlling Forgetting with Test-Time Data in Continual Learning&atitle=Controlling%20Forgetting%20with%20Test-Time%20Data%20in%20Continual%20Learning&aulast=Singh%2C%20Vaibhav&id=DOI: Name: Full Text Finder (for New FTF UI) (s8985755) Category: fullText Text: Find It @ SCU Libraries MouseOverText: Find It @ SCU Libraries |
|---|---|
| Header | DbId: edsarx DbLabel: arXiv An: edsarx.2406.13653 RelevancyScore: 1098 AccessLevel: 3 PubType: Report PubTypeId: report PreciseRelevancyScore: 1098.04455566406 |
| IllustrationInfo | |
| Items | – Name: Title Label: Title Group: Ti Data: Controlling Forgetting with Test-Time Data in Continual Learning – Name: Author Label: Authors Group: Au Data: <searchLink fieldCode="AR" term="%22Singh%2C+Vaibhav%22">Singh, Vaibhav</searchLink><br /><searchLink fieldCode="AR" term="%22Aljundi%2C+Rahaf%22">Aljundi, Rahaf</searchLink><br /><searchLink fieldCode="AR" term="%22Belilovsky%2C+Eugene%22">Belilovsky, Eugene</searchLink> – Name: DatePubCY Label: Publication Year Group: Date Data: 2024 – Name: Subset Label: Collection Group: HoldingsInfo Data: Computer Science – Name: Subject Label: Subject Terms Group: Su Data: <searchLink fieldCode="DE" term="%22Computer+Science+-+Machine+Learning%22">Computer Science - Machine Learning</searchLink> – Name: Abstract Label: Description Group: Ab Data: Foundational vision-language models have shown impressive performance on various downstream tasks. Yet, there is still a pressing need to update these models later as new tasks or domains become available. Ongoing Continual Learning (CL) research provides techniques to overcome catastrophic forgetting of previous information when new knowledge is acquired. To date, CL techniques focus only on the supervised training sessions. This results in significant forgetting yielding inferior performance to even the prior model zero shot performance. In this work, we argue that test-time data hold great information that can be leveraged in a self supervised manner to refresh the model's memory of previous learned tasks and hence greatly reduce forgetting at no extra labelling cost. We study how unsupervised data can be employed online to improve models' performance on prior tasks upon encountering representative samples. We propose a simple yet effective student-teacher model with gradient based sparse parameters updates and show significant performance improvements and reduction in forgetting, which could alleviate the role of an offline episodic memory/experience replay buffer.<br />Comment: 9 pages, 2 figures – Name: TypeDocument Label: Document Type Group: TypDoc Data: Working Paper – Name: URL Label: Access URL Group: URL Data: <link linkTarget="URL" linkTerm="http://arxiv.org/abs/2406.13653" linkWindow="_blank">http://arxiv.org/abs/2406.13653</link> – Name: AN Label: Accession Number Group: ID Data: edsarx.2406.13653 |
| PLink | https://login.libproxy.scu.edu/login?url=https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&scope=site&db=edsarx&AN=edsarx.2406.13653 |
| RecordInfo | BibRecord: BibEntity: Subjects: – SubjectFull: Computer Science - Machine Learning Type: general Titles: – TitleFull: Controlling Forgetting with Test-Time Data in Continual Learning Type: main BibRelationships: HasContributorRelationships: – PersonEntity: Name: NameFull: Singh, Vaibhav – PersonEntity: Name: NameFull: Aljundi, Rahaf – PersonEntity: Name: NameFull: Belilovsky, Eugene IsPartOfRelationships: – BibEntity: Dates: – D: 19 M: 06 Type: published Y: 2024 |
| ResultId | 1 |