报告题目:Learning to Communicate for few-shot learning
报 告 人:刘璐 博士(悉尼科技大学)
报告时间:2020年11月25日(周三)下午16:00-17:00
报告地点:4号楼4318室
邀 请 人:刘深泉 教授
欢迎广大师生前往!
数学学院
2020年11月24日
报告摘要:
Few-shot learning aims to train a model with only a small number of samples. In single domain few-shot classification problem, a class is underrepresented by few examples. For multi-domain few-shot classification problem, another challenge is how to transform a universal representation for all domains. In this talk, I will introduce how to communicate between classes and domains to improve the representation ability of models for both few-shot learning scenarios.