Learning with Limited Data
Yingjun Du (Ph.D. Student)
In clinical practices, massive annotations are difficult to acquire in some conditions where specialized biomedical expert knowledge is required. Inspired by the human ability to learn new tasks quickly from a small number of examples, few-shot learning attempts to address the challenge of training artiļ¬cial intelligence to generalize well with a small number of samples. Meta-learning addresses this problem by investigating common prior knowledge from previous tasks that can facilitate rapid learning of new tasks. However, due to the limited data, there are still many challenges with few-shot learning.
| Primary Advisor: | Cees Snoek (University of Amsterdam) |
| Industry Advisor: | Xiantong Zhen (University of Amsterdam & Inception Institute of Artificial Intelligence) |
| PhD Duration: | 01 June 2020 - 31 May 2024 |
