Customizing Student Networks From Heterogeneous Teachers via Adaptive Knowledge Amalgamation
arXiv:1908.07121
Abstract
A massive number of well-trained deep networks have been released by developers online. These networks may focus on different tasks and in many cases are optimized for different datasets. In this paper, we study how to exploit such heterogeneous pre-trained networks, known as teachers, so as to train a customized student network that tackles a set of selective tasks defined by the user. We assume no human annotations are available, and each teacher may be either single- or multi-task. To this end, we introduce a dual-step strategy that first extracts the task-specific knowledge from the heterogeneous teachers sharing the same sub-task, and then amalgamates the extracted knowledge to build the student network. To facilitate the training, we employ a selective learning scheme where, for each unlabelled sample, the student learns adaptively from only the teacher with the least prediction ambiguity. We evaluate the proposed approach on several datasets and experimental results demonstrate that the student, learned by such adaptive knowledge amalgamation, achieves performances even better than those of the teachers.
arXiv commentAccepted to ICCV 2019
References in corpus (8)
- Very Deep Convolutional Networks for Large-Scale Image Recognition
- Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
- Distilling the Knowledge in a Neural Network
- Paying More Attention to Attention: Improving the Performance of Convolutional Neural Networks via Attention Transfer
- Fine-Grained Visual Classification of Aircraft
- Like What You Like: Knowledge Distill via Neuron Selectivity Transfer
- Multi-Task Zipping via Layer-wise Neuron Sharing
- Knowledge Flow: Improve Upon Your Teachers