Analogical and Relational Reasoning with Spiking Neural Networks

Abstract

Raven's Progressive Matrices have been widely used for measuring abstract reasoning and intelligence in humans. However for artificial learning systems, abstract reasoning remains a challenging problem. In this paper we investigate how neural networks augmented with biologically inspired spiking modules gain a significant advantage in solving this problem. To illustrate this, we first investigate the performance of our networks with supervised learning, then with unsupervised learning. Experiments on the RAVEN dataset show that the overall accuracy of our supervised networks surpass human-level performance, while our unsupervised networks significantly outperform existing unsupervised methods. Finally, our results from both supervised and unsupervised learning illustrate that, unlike their non-augmented counterparts, networks with spiking modules are able to extract and encode temporal features without any explicit instruction, do not heavily rely on training data, and generalise more readily to new problems. In summary, the results reported here indicate that artificial neural networks with spiking modules are well suited to solving abstract reasoning.

arXiv commentProblems were discovered with the details of the experiments involving the LSM neural network, the results reported were not correct. A new version of the paper is in progress with corrected results

References in corpus (7)

• Adam: A Method for Stochastic Optimization
• Deep Neural Networks Rival the Representation of Primate IT Cortex for Core Visual Object Recognition
• Measuring abstract reasoning in neural networks
• Unsupervised Learning with Self-Organizing Spiking Neural Networks
• Improving Generalization for Abstract Reasoning Tasks Using Disentangled Feature Representations
• Learning Perceptual Inference by Contrasting
• Solving Raven's Progressive Matrices with Neural Networks