Back to Search Start Over

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

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

Authors :
Park, Haekyu
Lee, Seongmin
Hoover, Benjamin
Wright, Austin P.
Shaikh, Omar
Duggal, Rahul
Das, Nilaksh
Li, Kevin
Hoffman, Judy
Chau, Duen Horng
Park, Haekyu
Lee, Seongmin
Hoover, Benjamin
Wright, Austin P.
Shaikh, Omar
Duggal, Rahul
Das, Nilaksh
Li, Kevin
Hoffman, Judy
Chau, Duen Horng
Publication Year :
2022

Abstract

We present ConceptEvo, a unified interpretation framework for deep neural networks (DNNs) that reveals the inception and evolution of learned concepts during training. Our work addresses a critical gap in DNN interpretation research, as existing methods primarily focus on post-training interpretation. ConceptEvo introduces two novel technical contributions: (1) an algorithm that generates a unified semantic space, enabling side-by-side comparison of different models during training, and (2) an algorithm that discovers and quantifies important concept evolutions for class predictions. Through a large-scale human evaluation and quantitative experiments, we demonstrate that ConceptEvo successfully identifies concept evolutions across different models, which are not only comprehensible to humans but also crucial for class predictions. ConceptEvo is applicable to both modern DNN architectures, such as ConvNeXt, and classic DNNs, such as VGGs and InceptionV3.<br />Comment: Accepted at CIKM'23

Details

Database :
OAIster
Publication Type :
Electronic Resource
Accession number :
edsoai.on1333760808
Document Type :
Electronic Resource