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

Few-shot class-incremental learning (FSCIL), which targets at continuouslyexpanding model's representation capacity under few supervisions, is animportant yet challenging problem. On the one hand, when fitting new tasks(novel classes), features trained on old tasks (old classes) couldsignificantly drift, causing catastrophic forgetting. On the other hand,training the large amount of model parameters with few-shot novel-classexamples leads to model over-fitting. In this paper, we propose a learnableexpansion-and-compression network (LEC-Net), with the aim to simultaneouslysolve catastrophic forgetting and model over-fitting problems in a unifiedframework. By tentatively expanding network nodes, LEC-Net enlarges therepresentation capacity of features, alleviating feature drift of old networkfrom the perspective of model regularization. By compressing the expandednetwork nodes, LEC-Net purses minimal increase of model parameters, alleviatingover-fitting of the expanded network from a perspective of compactrepresentation. Experiments on the CUB/CIFAR-100 datasets show that LEC-Netimproves the baseline by 5~7% while outperforms the state-of-the-art by 5~6%.LEC-Net also demonstrates the potential to be a general incremental learningapproach with dynamic model expansion capability.

Learnable Expansion-and-Compression Network for Few-shot Class-Incremental Learning