Timezone: »

Learning Multiple Tasks with Multilinear Relationship Networks
Mingsheng Long · ZHANGJIE CAO · Jianmin Wang · Philip S Yu

Mon Dec 04 06:30 PM -- 10:30 PM (PST) @ Pacific Ballroom #7

Deep networks trained on large-scale data can learn transferable features to promote learning multiple tasks. Since deep features eventually transition from general to specific along deep networks, a fundamental problem of multi-task learning is how to exploit the task relatedness underlying parameter tensors and improve feature transferability in the multiple task-specific layers. This paper presents Multilinear Relationship Networks (MRN) that discover the task relationships based on novel tensor normal priors over parameter tensors of multiple task-specific layers in deep convolutional networks. By jointly learning transferable features and multilinear relationships of tasks and features, MRN is able to alleviate the dilemma of negative-transfer in the feature layers and under-transfer in the classifier layer. Experiments show that MRN yields state-of-the-art results on three multi-task learning datasets.

Author Information

Mingsheng Long (Tsinghua University)
ZHANGJIE CAO (Stanford University)
Jianmin Wang (Tsinghua University)
Philip S Yu (UIC)

More from the Same Authors