Three Analog Neurons Are Turing Universal
The languages accepted online by binary-state neural networks with rational weights have been shown to be context-sensitive when an extra analog neuron is added (1ANNs). In this paper, we provide an upper bound on the number of additional analog units to achieve Turing universality. We prove that any Turing machine can be simulated by a binary-state neural network extended with three analog neurons (3ANNs) having rational weights, with a linear-time overhead. Thus, the languages accepted offline by 3ANNs with rational weights are recursively enumerable, which refines the classification of neural networks within the Chomsky hierarchy.
KeywordsNeural computing Turing machine Chomsky hierarchy
- 7.Lupanov, O.B.: On the synthesis of threshold circuits. Probl. Kibern. 26, 109–140 (1973)Google Scholar
- 15.Šíma, J.: Analog stable simulation of discrete neural networks. Neural Netw. World 7(6), 679–686 (1997)Google Scholar
- 18.Šíma, J.: Neural networks between integer and rational weights. In: Proceedings of the IJCNN 2017 Thirties International Joint Conference on Neural Networks, pp. 154–161. IEEE (2017)Google Scholar