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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Continual learning aims to learn a sequence of tasks from dynamic datadistributions. Without accessing to the old training samples, knowledgetransfer from the old tasks to each new task is difficult to determine, whichmight be either positive or negative. If the old knowledge interferes with thelearning of a new task, i.e., the forward knowledge transfer is negative, thenprecisely remembering the old tasks will further aggravate the interference,thus decreasing the performance of continual learning. By contrast, biologicalneural networks can actively forget the old knowledge that conflicts with thelearning of a new experience, through regulating the learning-triggeredsynaptic expansion and synaptic convergence. Inspired by the biological activeforgetting, we propose to actively forget the old knowledge that limits thelearning of new tasks to benefit continual learning. Under the framework ofBayesian continual learning, we develop a novel approach named ActiveForgetting with synaptic Expansion-Convergence (AFEC). Our method dynamicallyexpands parameters to learn each new task and then selectively combines them,which is formally consistent with the underlying mechanism of biological activeforgetting. We extensively evaluate AFEC on a variety of continual learningbenchmarks, including CIFAR-10 regression tasks, visual classification tasksand Atari reinforcement tasks, where AFEC effectively improves the learning ofnew tasks and achieves the state-of-the-art performance in a plug-and-play way.

AFEC: Active Forgetting of Negative Transfer in Continual Learning