Incrementally Learned Angular Representations for Few-Shot Class-Incremental Learning

The main challenge of FSCIL is the trade-off between underfitting to a new session task and preventing forgetting the knowledge for earlier sessions. In this paper, we reveal that the angular space occupied by the features within the embedded area is relatively narrow. Consequently, after the base s...

Fuld beskrivelse

Bibliografiske detaljer

Principais autores:	In-Ug Yoon, Jong-Hwan Kim
Format:	Artigo
Sprog:	Inglês
Udgivet:	IEEE 2023-01-01
Serier:	IEEE Access
Fag:	Class-incremental learning few-shot learning few-shot class-incremental learning image classification
Online adgang:	https://ieeexplore.ieee.org/document/10353956/
Tags:	Tilføj Tag Ingen Tags, Vær først til at tagge denne postø!