- Student First Author: Yes
- Keywords: lifelong learning, continual learning
- Abstract: Effective lifelong learning across diverse tasks requires diverse knowledge, yet transferring irrelevant knowledge may lead to interference and catastrophic forgetting. In deep networks, transferring the appropriate granularity of knowledge is as important as the transfer mechanism, and must be driven by the relationships among tasks. We first show that the lifelong learning performance of several current deep learning architectures can be significantly improved by transfer at the appropriate layers. We then develop an expectation-maximization (EM) method to automatically select the appropriate transfer configuration and optimize the task network weights. This EM-based selective transfer is highly effective, as demonstrated on three algorithms in several lifelong object classification scenarios.
- TL;DR: This work empirically investigated the importance of transfer at the appropriate layers of deep neural network, and proposed EM-based selective transfer algorithm, which enhanced performance of several lifelong learning architectures.