.png){kind=icon}
.png){kind=icon}
Return
PSL: Prototype swapping learning for modality imbalance
.png){kind=icon}
PRE
AI
DOI：10.1016/j.neunet.2025.108322.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• PSLaddresses modality imbalance by swapping prototypes, exempting from imbalance metrics. • Hybrid prototype combines anchored and other modality prototypes for flexible manipulation. • PSLoutperforms the modulation methods based on the fusion training paradigm. • PSLachieves SoTA performance with an alternate training scheme integration.
Journal
IF:
6.3
Papers:
7.7K
Citations:
3.0W