Few-shot learner parameterization by diffusion time-steps

• Main Authors: YUE, Zhongqi, ZHOU, Pan, HONG, Richang, ZHANG, Hanwang, SUN Qianru
• 格式: text
• 語言: English
• 出版: Institutional Knowledge at Singapore Management University 2024
• 主題: Graphics and Human Computer Interfaces
• 在線閱讀: https://ink.library.smu.edu.sg/sis_research/9019; https://ink.library.smu.edu.sg/context/sis_research/article/10022/viewcontent/2024_CVPR_few_shot.pdf
• 機構: Singapore Management University

Even when using large multi-modal foundation models, few-shot learning is still challenging—if there is no proper inductive bias, it is nearly impossible to keep the nuanced class attributes while removing the visually prominent attributes that spuriously correlate with class labels. To this end, we...