Enhancing network modularity to mitigate catastrophic forgetting
- PDF / 1,991,010 Bytes
- 24 Pages / 595.276 x 790.866 pts Page_size
- 35 Downloads / 199 Views
pplied Network Science
Open Access
RESEARCH
Enhancing network modularity to mitigate catastrophic forgetting Lu Chen1* and Masayuki Murata2 *Correspondence: l‑[email protected] 1 Department of Information and Human Science, Kyoto Institute of Technology, Kyoto, Japan Full list of author information is available at the end of the article
Abstract Catastrophic forgetting occurs when learning algorithms change connections used to encode previously acquired skills to learn a new skill. Recently, a modular approach for neural networks was deemed necessary as learning problems grow in scale and complexity since it intuitively should reduce learning interference by separating functionality into physically distinct network modules. However, an algorithmic approach is difficult in practice since it involves expert design and trial and error. Kashtan et al. finds that evolution under an environment that changes in a modular fashion leads to the spontaneous evolution of a modular network structure. In this paper, we aim to solve the reverse problem of modularly varying goal (MVG) to obtain a highly modular structure that can mitigate catastrophic forgetting so that it can also apply to realistic data. First, we confirm that a configuration with a highly modular structure exists by applying an MVG against a realistic dataset and confirm that this neural network can mitigate catastrophic forgetting. Next, we solve the reverse problem, that is, we propose a method that can obtain a highly modular structure able to mitigate catastrophic forgetting. Since the MVG-obtained neural network can relatively maintain the intra-module elements while leaving the inter-module elements relatively variable, we propose a method to restrict the inter-module weight elements so that they can be relatively variable against the intra-module ones. From the results, the obtained neural network has a highly modular structure and can learn an unlearned goal faster than without this method. Keywords: Modularity, Catastrophic forgetting, Neural network
Introduction Learning a variety of different skills for different problems is a long-standing goal in artificial intelligence. However, when neural networks learn a new skill, they typically lose previously acquired skills. This problem, called catastrophic forgetting, occurs because learning algorithms change connections used to encode previously acquired skills to learn a new skill (Ellefsen et al. 2015; Goodfellow et al. 2014; Kemker et al. 2018). Catastrophic forgetting has been studied for a few decades (French 1999; Lee et al. 2017). Recently, a modular approach for neural networks has been deemed necessary as learning problems grow in scale and complexity (Amer and Maul 2019). This modular approach, which involves constructing a network with densely connected modules with only sparser connections between the modules (Newman 2006), intuitively should © The Author(s) 2020. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharin
Data Loading...
