Enabling Asymmetric Knowledge Transfer in Multi-Task Learning with Self-Auxiliaries
This addresses a specific bottleneck in multi-task learning for computer vision researchers, offering a novel optimization strategy to handle asymmetric task interactions.
The paper tackled the problem of asymmetric task relationships in multi-task learning, where knowledge transfer helps some tasks but hurts others, and proposed a method using self-auxiliaries to exploit this asymmetry, resulting in substantial performance improvements on benchmark computer vision problems.
Knowledge transfer in multi-task learning is typically viewed as a dichotomy; positive transfer, which improves the performance of all tasks, or negative transfer, which hinders the performance of all tasks. In this paper, we investigate the understudied problem of asymmetric task relationships, where knowledge transfer aids the learning of certain tasks while hindering the learning of others. We propose an optimisation strategy that includes additional cloned tasks named self-auxiliaries into the learning process to flexibly transfer knowledge between tasks asymmetrically. Our method can exploit asymmetric task relationships, benefiting from the positive transfer component while avoiding the negative transfer component. We demonstrate that asymmetric knowledge transfer provides substantial improvements in performance compared to existing multi-task optimisation strategies on benchmark computer vision problems.