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

In Class-Incremental Learning (CIL) an image classification system is exposedto new classes in each learning session and must be updated incrementally.Methods approaching this problem have updated both the classification head andthe feature extractor body at each session of CIL. In this work, we develop abaseline method, First Session Adaptation (FSA), that sheds light on theefficacy of existing CIL approaches and allows us to assess the relativeperformance contributions from head and body adaption. FSA adapts a pre-trainedneural network body only on the first learning session and fixes it thereafter;a head based on linear discriminant analysis (LDA), is then placed on top ofthe adapted body, allowing exact updates through CIL. FSA is replay-freei.e.~it does not memorize examples from previous sessions of continuallearning. To empirically motivate FSA, we first consider a diverse selection of22 image-classification datasets, evaluating different heads and bodyadaptation techniques in high/low-shot offline settings. We find that the LDAhead performs well and supports CIL out-of-the-box. We also find thatFeaturewise Layer Modulation (FiLM) adapters are highly effective in thefew-shot setting, and full-body adaption in the high-shot setting. Second, weempirically investigate various CIL settings including high-shot CIL andfew-shot CIL, including settings that have previously been used in theliterature. We show that FSA significantly improves over the state-of-the-artin 15 of the 16 settings considered. FSA with FiLM adapters is especiallyperformant in the few-shot setting. These results indicate that currentapproaches to continuous body adaptation are not working as expected. Finally,we propose a measure that can be applied to a set of unlabelled inputs which ispredictive of the benefits of body adaptation.

First Session Adaptation: A Strong Replay-Free Baseline for Class-Incremental Learning