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

Modern computer vision applications suffer from catastrophic forgetting whenincrementally learning new concepts over time. The most successful approachesto alleviate this forgetting require extensive replay of previously seen data,which is problematic when memory constraints or data legality concerns exist.In this work, we consider the high-impact problem of Data-FreeClass-Incremental Learning (DFCIL), where an incremental learning agent mustlearn new concepts over time without storing generators or training data frompast tasks. One approach for DFCIL is to replay synthetic images produced byinverting a frozen copy of the learner's classification model, but we show thisapproach fails for common class-incremental benchmarks when using standarddistillation strategies. We diagnose the cause of this failure and propose anovel incremental distillation strategy for DFCIL, contributing a modifiedcross-entropy training and importance-weighted feature distillation, and showthat our method results in up to a 25.1% increase in final task accuracy(absolute difference) compared to SOTA DFCIL methods for commonclass-incremental benchmarks. Our method even outperforms several standardreplay based methods which store a coreset of images.

Always Be Dreaming: A New Approach for Data-Free Class-Incremental Learning