350 rub
Journal Neurocomputers №12 for 2016 г.
Article in number:
On a Predictive Model of Intelligence
Authors:
M.A. Sushchin - Ph.D. (Phil.), Senior Research Scientist, Department of Philosophy and Sociology of Southwest State University (Kursk); Institute of Scientific Information of Social Sciences of the Russian Academy of Sciences (Moscow)
E-mail: sushchin@bk.ru
Abstract:
This article treats Jeff Hawkins - memory-prediction framework of intelligence (worked out with his co-author Sandra Blakeslee) in an unusual way. Instead of simply considering its virtues and perils, the author tries to compare Hawkins - model of in-telligence to a very different tradition in contemporary cognitive investigations, known primarily as situated and embodied cognition. The author argues that in spite of a number of opposite philosophical presuppositions and quite a different research focus, these two frameworks share a lot of profoundly similar claims about the nature of cognitive processes.
Among these deep similarities, according to the author, the most important is probably an appreciation of the central role of context and current situation in the workings of intelligence and cognition. Thus, on Hawkins - approach the brain-s crucial ability to use invariant memories to make sensorimotor predictions about what the organism will experience next is only possible by means of combining those invariant memory recalls with the details pertaining to every new unique situation. In this respect, the author points out that a very similar emphasis upon the immense importance of the context was one of the pillars of the entire situated cognition movement. Besides that, the two frameworks are characterized by a very similar assertion about close couplings between perceptual and motor facets of cognitive activity.
Furthermore, the author suggests that the key thesis of contemporary enactivism - that perception requires an understanding of sensorimotor dependencies - needs to be assimilated by the kind of predictive theory of cognition and intelligence Hawkins proposes in his book. From the point of view of predictive framework, this simply means that our understanding of sensorimotor dependencies, when we know the effects of our bodily movements on sensory stimulation, is nothing else than ability to make sensorimotor predictions on the basis of past experience and memories stored in our brains. The author argues that with this appreciation of the central role of the memory-based predictive activity of the brain Hawkins - theory is able to fill in "the explanatory gap" inherent to antirepresentational approaches to perception and cognition like that of J.J. Gibson and modern enactivists.
In the central part of the work the author explores a source of a potential theoretical conflict between the two theories which is caused by their different conceptualization of memory and its role in the cognitive evolution of man. Thus, Hawkins - proposal is centered on a wholly biological sense of memory as an ultimate goal of the evolution of intelligent agents culminating in a complex human neocortical memory system whereas theorists of situated and embodied tradition in cognitive science rested heavily on a radical idea of the world as an outside memory (or the world as its own best model, as it was put by a roboticist Rodney Brooks). For them this ability to complement and enhance our basic internal resources by creating and exploiting various kinds of intelligent environments was perhaps the most distinctive cognitive achievement of the evolution of man. Nevertheless the author concludes that these two opposite views of memory can be reconciled and integrated into a coherent perspective on the nature and evolution of intelligence. Biological approach to memory might be correct in that it captures one of the general trends in the evolution of intelligent systems on a large scale whereas the externalist approach is acceptable because it is focused on the most recent revolutionary achievement of humans which has become available to them on later stages of anthropogenesis.
The author wrote this review partly because he shares Hawkins - deep conviction that without a unified theory of mind/brain we will be unable to build truly intelligent machines and artifacts, a methodology that was recently taken over by the field of reverse engineering of the brain. Based on several recent discussions of Bayesian approaches, the author concludes that though these projects may be an important step towards a unified theory in cognitive science and neuroscience, in their present form they lack a crucial body of evidence so that we can adjudicate conclusively whether they are satisfactory or not. This does not mean that we are hostages to blind processes of accretion of empirical details, out of which a unified theory of mind/brain will somehow emerge. According to the author of this article, this simply means that current theoretical approaches in cognitive science need much stronger confirmation to be able to reveal the mysteries of intelligence and the brain.
Pages: 31-39
References
- The Future of the Brain: Essays by the World-s Leading Neuroscientists / Ed. by G. Marcus, J. Freeman. Princeton: Princeton University Press. 2014.
- Arshinov V.I., Aseeva I.A., Budanov V.G., Grebenshhikova E.G., Grimov O.A., Kamenskijj E.G., Moskalev I.E., Pirozhkova S.V., Sushhin M.A., CHeklecov V.V. Socio-antropologicheskie izmerenija konvergentnykh tekhnologijj (materialy «kruglogo stola») // Filosofskie nauki. 2015. № 11. S. 135-147.
- Hohwy J. The Predictive Mind. New York: Oxford University Press. 2013.
- Clark A. Whatever next - Predictive brains, situated agents, and the future of cognitive science // Behavioral and Brain Sciences. 2013. V. 36. № 3. P. 181-204.
- Hawkins J., Blakeslee S. On Intelligence. New York: An Owl Book. 2007.
- Hinton G. E. Learning multiple layers of representation // Trends in Cognitive Sciences. 2007. V. 11. № 10. P. 428-434.
- Azevedo F.C. A. et al. Equal Numbers of Neuronal and Nonneuronal Cells Make the Human Brain an Isometrically Scaled-Up Primate Brain // The Journal of Comparative Neurology. 2009. V. 513. № 5. P. 532-541.
- Brooks R. Cambrian Intelligence: The Early History of the New AI. Cambridge, MA: A Bradford Book/The MIT Press. 1999.
- Hutchins E. Cognition in the Wild. Cambridge, MA: A Bradford Book/The MIT Press. 1995.
- Clark A., Chalmers D. The Extended Mind // Analysis. 1998. V. 58. № 1. P. 7-19.
- Van Gelder T. J. What Might Cognition Be, If Not Computation - // The Journal of Philosophy. 1995. V. 92. № 7. P. 345-381.
- Clark A. Being There: Putting Brain, Body and World Together Again. Cambridge, MA: A Bradford Book/The MIT Press. 1998.
- Noe A. Action in Perception. Cambridge. MA: The MIT Press. 2004.
- Donald M. Précis of Origins of the Modern Mind: Three Stages in Evolution of Culture and Cognition // Behavioral and Brain Sciences. 1993. V. 16. № 4. P. 737-748.
- Bowers J. S., Davis C. J. Bayesian Just-So-Stories in Psychology and Neuroscience // Psychological Bulletin. 2012. V. 138. № 3. P. 389-414.
- Marr D. Vision. San Francisco: Freeman. 1982.