English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Continual learning of new knowledge over time is one desirable capability forintelligent systems to recognize more and more classes of objects. Without orwith very limited amount of old data stored, an intelligent system oftencatastrophically forgets previously learned old knowledge when learning newknowledge. Recently, various approaches have been proposed to alleviate thecatastrophic forgetting issue. However, old knowledge learned earlier iscommonly less preserved than that learned more recently. In order to reduce theforgetting of particularly earlier learned old knowledge and improve theoverall continual learning performance, we propose a simple yet effectivefusion mechanism by including all the previously learned feature extractorsinto the intelligent model. In addition, a new feature extractor is included tothe model when learning a new set of classes each time, and a feature extractorpruning is also applied to prevent the whole model size from growing rapidly.Experiments on multiple classification tasks show that the proposed approachcan effectively reduce the forgetting of old knowledge, achievingstate-of-the-art continual learning performance.

Preserving Earlier Knowledge in Continual Learning with the Help of All Previous Feature Extractors