Schizophrenia and the Future of Artificial Intelligence
In the Meaningful-Based Cognitive Architecture (MBCA) the input sensory vector is propagated through a hierarchy of Hopfield-like Network (HLN) functional groups, including a binding of sensory input group of HLNs and a causal group of HLNs, and subsymbolic processing of the input vector occurs in the process. However, the processed sensory input vector is also propagated to the logic/working memory groups of HLNs, where the content of the logic/working memory can be compared to data held by various groups of HLNs of other functional groups, as well as with other logic/working memory units, i.e., symbolic processing occurs. While the MBCA does not attempt to replicate biological systems at the neuronal spiking level, its HLNs and the organization of its HLNs are indeed inspired by biological mammalian minicolumns and mammalian brains. The MBCA model leads to the hypothesis that in the course of hominid evolution, HLNs became co-opted into groups of HLNs providing more extensive working memories with more logical abilities. While such co-option of the minicolumns can allow advantageous symbolic processing integrated with subsymbolic processing, the order of magnitude of increased complexity required for such organization and operation, created a vulnerability in the human brain to psychotic disorders. The emergence of a technological artificial general intelligence (AGI) will, on a practical level, also require the integration of symbolic processing along with subsymbolic structures. The MBCA model predicts that such an integration will similarly create a potential vulnerability in the resultant AGI towards psychotic-like features.
KeywordsCognitive architecture Artificial general intelligence Cortical minicolumns Psychosis
This article builds upon work originally presented at BICA 2018 (reference 1).
- 2.Schneider H (2019, in press) Subsymbolic versus symbolic data flow in the meaningful-based cognitive architecture. In: Samsonovich AV (ed) Biologically inspired cognitive architectures 2019. Advances in intelligent systems and computing. SpringerGoogle Scholar
- 9.Bach J (2008) Seven principles of synthetic intelligence. In: Proceedings of 1st conference on artificial general intelligence, Memphis, pp. 63–74Google Scholar
- 12.Collier M, Beel J (2018) Implementing neural Turing machines. In: 27th international conference on artificial neural networks (ICANN), Rhodes, Greece, 4–7 OctoberGoogle Scholar
- 13.Li S, Xu H, Lu Z (2018) Generalize symbolic knowledge with neural rule engine. arXiv:1808.10326
- 14.Lázaro-Gredilla M, Liu Y, Phoenix DS, George D (2017) Hierarchical compositional feature learning. arXiv preprint. arXiv:1611.02252v2
- 15.Hawkins J, Blakeslee S (2004) On intelligence. Times Books, New YorkGoogle Scholar
- 16.Kurzweil R (2012) How to create a mind. Viking Press, New YorkGoogle Scholar
- 17.Sabour S, Frosst N, Hinton GE (2017) Dynamic routing between capsules. In: Conference on neural information processing systems (NIPS 2017), Long Beach, CA, USAGoogle Scholar
- 18.Schneider H (2019 in press) Emergence of belief systems and the future of artificial intelligence. In: Samsonovich AV (ed) Biologically inspired cognitive architectures 2019. Advances in intelligent systems and computing. SpringerGoogle Scholar
- 28.Anttila V, Bulik-Sullivan B et al (2018) Analysis of shared heritability in common disorders of the brain. Science 360(6395):eaap8757. https://doi.org/10.1126/science.aap8757