How important is the train-validation split in meta-learning?

• Main Authors: BAI, Yu, CHEN, Minshuo, ZHOU, Pan, ZHAO, Tuo, LEE, D. Jason, KAKADE, Sham, WANG, Huan, XIONG, Caiming
• Format: text
• Language: English
• Published: Institutional Knowledge at Singapore Management University 2021
• Subjects: Artificial Intelligence and Robotics; Graphics and Human Computer Interfaces
• Online Access: https://ink.library.smu.edu.sg/sis_research/8991; https://ink.library.smu.edu.sg/context/sis_research/article/9994/viewcontent/2021_ICML_Metalearning.pdf
• Institution: Singapore Management University

Meta-learning aims to perform fast adaptation on a new task through learning a “prior” from multiple existing tasks. A common practice in meta-learning is to perform a train-validation split (train-val method) where the prior adapts to the task on one split of the data, and the resulting predictor i...