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 faces a crucial challenge of catastrophic forgetting. Toaddress this challenge, experience replay (ER) that maintains a tiny subset ofsamples from previous tasks has been commonly used. Existing ER works usuallyfocus on refining the learning objective for each task with a static memoryconstruction policy. In this paper, we formulate the dynamic memoryconstruction in ER as a combinatorial optimization problem, which aims atdirectly minimizing the global loss across all experienced tasks. We firstapply three tactics to solve the problem in the offline setting as a startingpoint. To provide an approximate solution to this problem in the onlinecontinual learning setting, we further propose the Global Pseudo-taskSimulation (GPS), which mimics future catastrophic forgetting of the currenttask by permutation. Our empirical results and analyses suggest that the GPSconsistently improves accuracy across four commonly used vision benchmarks. Wehave also shown that our GPS can serve as the unified framework for integratingvarious memory construction policies in existing ER works.

Navigating Memory Construction by Global Pseudo-Task Simulation for Continual Learning