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

We present ConceptEvo, a unified interpretation framework for deep neuralnetworks (DNNs) that reveals the inception and evolution of learned conceptsduring training. Our work addresses a critical gap in DNN interpretationresearch, as existing methods primarily focus on post-training interpretation.ConceptEvo introduces two novel technical contributions: (1) an algorithm thatgenerates a unified semantic space, enabling side-by-side comparison ofdifferent models during training, and (2) an algorithm that discovers andquantifies important concept evolutions for class predictions. Through alarge-scale human evaluation and quantitative experiments, we demonstrate thatConceptEvo successfully identifies concept evolutions across different models,which are not only comprehensible to humans but also crucial for classpredictions. ConceptEvo is applicable to both modern DNN architectures, such asConvNeXt, and classic DNNs, such as VGGs and InceptionV3.

Concept Evolution in Deep Learning Training: A Unified Interpretation Framework and Discoveries