2022 (modified: 23 Jan 2023)ICML 2022Readers: Everyone
Abstract:Existing auxiliary learning approaches only consider the relationships between the target task and the auxiliary tasks, ignoring the fact that data samples within an auxiliary task could contribute...