Dendritic Self-Organizing Maps for Continual Learning

Abstract

Current deep learning architectures show remarkable performance when trained in large-scale, controlled datasets. However, the predictive ability of these architectures significantly decreases when learning new classes incrementally. This is due to their inclination to forget the knowledge acquired from previously seen data, a phenomenon termed catastrophic-forgetting. On the other hand, Self-Organizing Maps (SOMs) can model the input space utilizing constrained k-means and thus maintain past knowledge. Here, we propose a novel algorithm inspired by biological neurons, termed Dendritic-Self-Organizing Map (DendSOM). DendSOM consists of a single layer of SOMs, which extract patterns from specific regions of the input space accompanied by a set of hit matrices, one per SOM, which estimate the association between units and labels. The best-matching unit of an input pattern is selected using the maximum cosine similarity rule, while the point-wise mutual information is employed for class inference. DendSOM performs unsupervised feature extraction as it does not use labels for targeted updating of the weights. It outperforms classical SOMs and several state-of-the-art continual learning algorithms on benchmark datasets, such as the Split-MNIST and Split-CIFAR-10. We propose that the incorporation of neuronal properties in SOMs may help remedy catastrophic forgetting.

Introducing the Dendrify framework for incorporating dendrites to spiking neural networks

“…Similar advantages were recently seen in the machine learning field: adding dendritic nodes in artificial neural networks (ANNs) reduced the number of trainable parameters required to achieve high-performance accuracy 37 (also see 38,39 ). Moreover, incorporating dendritic nodes in Self Organizing Map classifiers 40 and other neuro-inspired networks 41 improved their continuous learning ability.…”

Introducing the Dendrify framework for incorporating dendrites to spiking neural networks

“…Similar advantages were recently seen in the machine learning field: adding dendritic nodes in artificial neural networks (ANNs) reduced the number of trainable parameters required to achieve highperformance accuracy 37 (also see 38,39 ). Moreover, incorporating dendritic nodes in Self Organizing Map classifiers 40 and other neuroinspired networks 41 improved their continuous learning ability.…”