A Unified Hierarchy for AI and Natural Intelligence through Auto-Programming for General Purposes
0
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
Despite the current great public interest in neural network based artificial intelligence, there exists a huge gap between artificial intelligence (AI) and natural intelligence (NI). Furthermore, there is such a lack of unified hierarchy of intelligence that intelligence is widely regarded piecemeal. This status quo results in highly brittle AI systems. This paper proposes how an autonomous agent, natural or artificial, develops a unified intelligence hierarchy in the brain. The term “unified” means not only for AI and NI both, but also for all practical sensory modalities and motor modalities, including perception, representation, reasoning, learning, societal activities and politics. This line of work has been supported by rigorous mathematical proofs and initial experimental verifications. However, this paper minimizes mathematical material so that the new information here can reach a wide audience. We should ask a new and general question: “How can a machine, natural or artificial, Autonomously Program For General Purposes (APFGP) from the real physical world?” We have given this question a solution, theoretically, experimentally, and mathematically. A clear but powerful learning engine—Developmental Network (DN) enables APFGP. Hopefully, understanding APFGP for both AI and NI not only fully automates development of AI systems but also improve human development, individually and societally.
Abstract
Despite the current great public interest in neural network based artificial intelligence, there exists a huge gap between artificial intelligence (AI) and natural intelligence (NI). Furthermore, there is such a lack of unified hierarchy of intelligence that intelligence is widely regarded piecemeal. This status quo results in highly brittle AI systems. This paper proposes how an autonomous agent, natural or artificial, develops a unified intelligence hierarchy in the brain. The term “unified” means not only for AI and NI both, but also for all practical sensory modalities and motor modalities, including perception, representation, reasoning, learning, societal activities and politics. This line of work has been supported by rigorous mathematical proofs and initial experimental verifications. However, this paper minimizes mathematical material so that the new information here can reach a wide audience. We should ask a new and general question: “How can a machine, natural or artificial, Autonomously Program For General Purposes (APFGP) from the real physical world?” We have given this question a solution, theoretically, experimentally, and mathematically. A clear but powerful learning engine—Developmental Network (DN) enables APFGP. Hopefully, understanding APFGP for both AI and NI not only fully automates development of AI systems but also improve human development, individually and societally.
참고문헌 (53)
- Carpenter, G. A., "ARTMAP: Supervised real-time learning and classification of nonstationary data by a self-organizing neural networks", Neural Networks, 4, pp.565~588 (1991)
- Chomsky, "Three models for the description of language", IRE Transactions on Information Theory, 2, pp.113~124 (1956)
- Chomsky, "On certain formal properties of grammars", Information and Control, 2(2), pp.137~167 (1959)
- Chomsky, N., "Computer Programming and Formal Languages", pp.118~161 (1963), North Holland
- Daw, N.D., "Opponent interactions between serotonin and dopamine", Neural Networks, 15(4-6), pp.603~616 (2002)
- Fukushima, K., "Cognitron: A self-organizing multilayered neural network", Biological Cybernetics, 20, pp.121~136 (1975)
- Graves, A., "Hybrid computing using a neural network with dynamic external memory", Nature, 538, pp.471~476 (2016)
- Guo, Q., "Cross-domain and within-domain synaptic maintenance for autonomous development of visual areas", Proc. the Fifth Joint IEEE International Conference on Development and Learning and on Epigenetic Robotics, pp.1~6 (2015)
- Hinton, G. E., "A fast learning algorithm for deep belief nets", Neural Computation, 18, pp.1527~1554 (2006)
- Ji, Z., "Where-What Network 1: “Where” and “What” assist each other through top-down connections", Proc. IEEE Int’l Conference on Development and Learning, pp.61~66 (2008)
- Jordan, M.I., "Machine learning: Trends, perspectives, and prospects", Science, 349, pp.255~260 (2015)
- Kakade, S., "Dopamine: generalization and bonuses", Neural Network, 15, pp.549~559 (2002)
- Krishna, R., "Visual genome", International Journal of Computer Vision, 123(1), pp.32~73 (2017)
- Krizhevsky, A., "Imagenet classification with deep convolutional neural networks", Advances in Neural Information Processing Systems, 25, pp.1106~1114 (2012)
- Laird, J. E., "An integration of external interaction, planning, and learning using Soar", Robotics and Autonomous Systems, 8, pp.113~129 (1991)
- Lake, B. M., "Human-level concept learning through probabilistic program induction", Science, 350, pp.1332~1338 (2016)
- LeCun, Y., "Deep learning", Nature, 521, pp.436~444 (2015)
- Marcus, G., "The Hapgazard Evolution of the Human Mind" (2009), Houghton Mifflin Harcourt
- Martin, J.C., "Introduction to Languages and the Theory of Computation" (2011), McGraw Hill
- Meehan, J.R., "The metanovel: Writing stories by computer", Ph. D., pp.1976
- Mnih, V., "Human-level control through deep reinforcement learning", Nature, 518, pp.529~533 (2015)
- Moran, J., "Selective attention gates visual processing in the extrastrate cortex", Science, 229(4715), pp.782~784 (1985)
- Moravcik, M., "Deepstack: Expert-level artificial intelligence in heads-up nolimit poker", Science, 356, pp.508~513 (2017)
- Olshausen, B. A., "A neurobiological model of visual attention and invariant pattern recognition based on dynamic routing of information", Journal of Neuroscience, 13(11), pp.4700~4719 (1993)
- Omlin, C. W., "Constructing deterministic finite-state automata in recurrent neural networks", Journal of the ACM, 43(6), pp.937~972 (1996)
- Rogers, T. T., "Précis of semantic cognition: A parallel distributed processing approach", Behavioral and Brain Sciences, 31, pp.689~749 (2008)
- Schmidhuber, J., "Deep learning in neural networks: An overview", Neural Networks, 61, pp.85~117 (2015)
- Scriven, R., "Collins French grammar" (2011), HarperCollins
- Siegelmann, H.T., "Computation beyond the Turing limit", Science, 286, pp.545~548 (1995)
- Solgi, M., "WWN-8: Incremental online stereo with shape-from-x using lifelong big data from multiple modalities", Proc. INNS Conference on Big Data, pp.316~326 (2015)
- Sun, R., "The interaction of the explicit and the implicit in skill learning: A dual-process approach", Psychological Review, 112(1), pp.59~192 (2005)
- Sutton, R.S., "Reinforcement Learning" (1998), MIT Press
- Wang, Y., "Synapse maintenance in the Where-What Network", Proc. Int’l Joint Conference on Neural Networks, pp.2823~2829 (2011)
- Weng, J., "Why have we passed “neural networks do not abstract well”?", Natural Intelligence: the INNS Magazine, 1(1), pp.13~22 (2011)
- Weng, J., "Natural and Artificial Intelligence: Introduction to Computational Brain-Mind" (2012), BMI Press
- Weng, J., "Symbolic models and emergent models: A review", IEEE Trans. Autonomous Mental Development, 4(1), pp.29~53 (2012)
- Weng, J., "How the brain-mind works: A two-page introduction to a theory", Brain-Mind Magazine, 2(2), pp.1~3 (2013)
- Weng, J., "Brain as an emergent finite automaton: A theory and three theorems", International Journal of Intelligent Science, 5(2), pp.112~131 (2015)
- A model for auto-programming for general purposes, "Weng, J.", Michigan State University, ~ (2017)
- Weng, J., "Natural and Artificial Intelligence: Introduction to Computational Brain-Mind" (2019), BMI Press
- Weng, J., "Learning recognition and segmentation of 3-D objects from 2-D images", Proc. IEEE 4th Int’l Conf. Computer Vision, pp.121~128 (1993)
- Weng, J., "Learning recognition and segmentation using the Cresceptron", International Journal of Computer Vision, 25(2), pp.109~143 (1997)
- Weng, J., "Transitory image sequences, asymptotic properties, and estimation of motion and structure", IEEE Trans. Pattern Analysis and Machine Intelligence, 19(5), pp.451~464 (1997)
- Weng, J., "Autonomous mental development by robots and animals", Science, 291(5504), pp.599~600 (2001)
- Weng, J., "Emergent Turing machines and operating systems for brain-like auto-programming for general purposes", Proc. AAAI 2018 Fall Symposium: Gathering for Artificial Intelligence and Natural Systems, pp.1~7 (2018)
- Weng, J., "Developmental network 2 (DN-2) and autoprogramming operating systems (AOS)", Patent document in preparation to be submitted to USPTO (2017), the data and in-house performance are being used in the AIML Contest 2017, ~ (2017)
- Weng, J., "Auto-programming for general purposes: Theory and experiments", Proc. International Joint Conference on Neural Networks, pp.1~8 (2020)
- Yu, A. J., "Uncertainty, neuromodulation, and attention", Neuron, 46, pp.681~692 (2005)
- Zheng, Z., "Emergent neural Turing machine and its visual navigation", Neural Networks, 110, pp.116~130 (2019)