浏览全部资源
扫码关注微信
网络出版日期： 2024-10-16 ，
移动端阅览
鲍泓,郑颖,梁天骄.认知机器的模型与结构研究进展[J].中国图象图形学报,
Bao Hong,Zheng Ying,Liang Tianjiao.Research Progress and Trends in Models and Structures of Cognitive Machines[J].Journal of Image and Graphics,
目的
2
机器如何同人一样具有认知能力，认知能力可用智能度量，人的智能是认知过程的涌现，我们从认知的模型出发研究其结构，结构决定机器的认知功能。本文旨在探讨机器认知的模型和构建方法，为设计新一代认知机器提供新的结构和方法论。
方法
2
本文用分析、归纳和演绎的方法综述认知机器模型和结构的起源、演进与发展趋势。首先，从20世纪初以来计算机器的发明和DNA双螺旋结构模型的发现谈起，阐述了“图灵机模型+冯·诺依曼结构”划时代意义的科学研究成果的形成，这一模型和结构催生了通用计算机器的发明，并对计算机科学与技术等新学科的形成起到奠基作用；此后，图灵的天问“机器能思维吗”及 “图灵测试”对后来建立人工智能有重大启示和影响；然后评述近二十年来 “深度学习模型+卷积神经网结构”以及 “大模型+转换器结构”等的里程碑式进展和存在的问题；在最新进展部分综述当前国内外有代表性的三位科学家提出的模型和结构：“世界模型”、“空间智能”和“认知螺旋”，特别是李德毅创立的认知物理学为机器认知提供了统一的理论框架，构成了机器认知的四种基本模式--认知螺旋结构模型和OOXA结构链，讨论了认知核、洋葱模型和负熵概念，以驾驶脑认知为案例进行实验验证；最后，展望了本领域未来研究和发展趋势。
结论
2
模型定义了机器思维的约束边界，结构决定机器的涌现性，通过模型+结构的研究方法和评价，为解决“机器如何像人一样思维”这样的人工智能重大问题提供了一种研究思路和范式。
How can machines possess cognitive abilities akin to humans？ The measure of cognitive ability is intelligence. Human intelligence is an emergent property of cognitive processes. We investigate the structure of cognition from cognitive models， as structure determines the cognitive functions of machines. The goal of this research is to provide innovative architectural and technological methodologies for the design of a new generation of intelligent machines. How can machines possess cognitive abilities similar to humans？ This requires starting from cognitive models to study the boundary constraints of human perception of the physical world and the mapping to cognitive space. The structure of the model determines its cognitive functions， and the combination of "model + structure" determines the emergent properties of the machine （system）. This paper utilizes a comprehensive approach that includes analysis， tracing to the source， and deduction to provide an in-depth review of the origins， the evolutionary path， and the emerging trends within the domain of cognitive machine models and their structural development. It begins by revisiting the cross-disciplinary interpretation of physicist Erwin Schrödinger's "thought experiment model and parallel universe view" since the early 20th century， which initiated a cognitive revolution with the concept that "life is information， living on negative entropy." This discovery not only unveiled the mysteries of life but also led to the emergence of new disciplines such as molecular biology and genomics. The computational theory， cybernetics， and information theory established during the same era continue to influence the development of information-physical systems to this day， with the "Turing machine model + von Neumann architecture" laying the foundation for the invention of general-purpose computing machines and the formation of new disciplines such as computer science and technology. Furthermore， Turing's profound question "Can machines think？" and his "Turing test" for measuring machine intelligence have attempted to unravel the mystery of the emergence of intelligence， having a significant inspiration and impact on the establishment of the artificial intelligence discipline. This paper focuses on analyzing and reviewing the milestone advancements and existing issues of the "Deep Learning Models + Convolutional Neural Networks" and the "Large Language Model + Transformer Structures" established over the past two decades. These issues also arise from several structural limitations when AI systems operate on computers with von Neumann architecture. To address these issues and overcome structural limitations， this paper reviews in the latest developments section the models and structures proposed by three of the most representative scientists both domestically and internationally： Yann LeCun's "World Model + Self-Supervision，" Fei-Fei Li's "Spatial Intelligence + Behavioral Vision，" and Deyi Li's "Cognitive Spiral + Cognitive Structural Chain." In particular， the cognitive physics established by Li Deyi provides a unified theoretical framework for machine cognition， constituting four basic patterns of machine cognition—the Cognitive Helix structural model and the OOXA structural chain. It indicating that there are four fundamental cognitive modes for humans： induction， deduction， creation， and discovery. Formalized as four basic modes of machine cognition： Observe Orient Act （OOA）， Observe Orient Decide Act（OODA）， Observe Orient Create Act（OOCA）， and Observe Orient Hypothesis Act （OOHA）. The paper discusses the concept of machines relying on negative entropy and its measurement methods， and taking our research and application of the cognitive process of unmanned driving machines as an example， it provides ideas， theoretical foundations， and architectural frameworks for further research and the establishment of credible， controllable， and interpretable machine cognition models and structures. Finally， this paper provides a prospective overview of future research and the development trends in this field.
Conclusion
2
Structure is the cornerstone of machine cognition， and "model + structure" determines the emergence of machines （systems）. The research methodology and evaluation of "model + structure" provide us with a research idea and paradigm for solving the major AI problem of "how to think like a human".
认知机器认知物理学认知核模型结构涌现负熵
cognition machinecognitive physicscognitive nucleusmodelstructureemergencenegentropy
Alan M .Turing. 1950. Computing machinery and intelligence. Mind， 59： 433-460.
Ashish Vaswani， Noam M.Shazeer， Niki Parmar， Jakob Uszkoreit， Llion Jones， Aidan N.Gomez， Lukasz Kaiser and Illia Polosukhin. 2017. Attention Is All You Need. Neural Information Processing Systems.［DOI：10.48550/arXiv.1706.03762http://dx.doi.org/10.48550/arXiv.1706.03762.］
Claude Elwood Shannon. 1948. A Mathematical Theory of Communication.
David E. Rumelhart， Geoffrey E. Hinton and Ronald J. Williams. 1986. Learning representations by back-propagating errors. Nature 323， 533–536. ［https：//doi.org/10.1038/323533a0https://doi.org/10.1038/323533a0］
David H. Ackley， Geoffrey E. Hinton and Terrence J. Sejnowski. 1985. A Learning Algorithm for Boltzmann Machines.Cognitive Science，9（1）：147-168
Di Kaichang， Li Deren and Li Deyi. 1998. Extension of Cloud Models and Their Applications in Spatial Data Mining. National Conference on Image and Graphics Technology. （邸凯昌，李德仁，李德毅. 1998： 530-535）.云模型的扩展及其在空间数据发掘中的应用.全国图象图形科技大会. ［DOI：ConferenceArticle/ 5aa7228cc095d72220f96400http://dx.doi.org/ConferenceArticle/5aa7228cc095d72220f96400.］
Erwin Schrodinger.1943. What is Life？ The Physical Aspect of the Living Cell， Cambridge of the University Press.
Gao Wang ， Giulia Marcucci， Benjamin Peters ，Maria Chiara Braidotti，Lars Muckli Daniele Faccio. Human-centred physical neuromorphics with visual brain-computer interfaces. nature communications. https：/doi.org/10.1038/s41467-024-50775-2https://doi.org/10.1038/s41467-024-50775-2
seceived： 14 November 2023， Accepted： 19 July 2024.
Hedrick J. K.， McMahon D.， Narendran V. and Swaroop D..1991. Longitudinal Vehicle Controller Design for IVHS Systems. American Control Conference， Boston， MA， USA. 3107-3112. ［ DOI： 10.23919/ACC.1991.4791980http://dx.doi.org/10.23919/ACC.1991.4791980］
Jascha Achterberg， Danyal Akarca， DJ Strouse， John Duncan and Duncan Astle. 2023. Spatially embedded recurrent neural networks reveal widespread links between structural and functional neuroscience findings. Nature Machine Intelligence， 5（12）： 1369-1381［DOI： 10.1038/s42256-023-00748-9http://dx.doi.org/10.1038/s42256-023-00748-9］
Johnjoe McFadden. 2020.Integrating information n the brain's EM field： the cemi field theory of consciousness. Neuroscience of Consciousness， Volume 2020， Issue 1.
Josh Abramson， Jonas Adler， Jack Dunger， Richard Evans， Tim Green， Alexander Pritzel， Olaf Ronneberger，Lindsay Willmore， Andrew J. Ballard， Joshua Bambrick， Sebastian W. Bodenstein， David A. Evans， Chia-Chun Hung， Michael O’Neill， David Reiman， Kathryn Tunyasuvunakool， Zachary Wu， Akvilė Žemgulytė， Eirini Arvaniti， Charles Beattie， Ottavia Bertolli， Alex Bridgland， Alexey Cherepanov， Miles Congreve， Alexander I. Cowen-Rivers， Andrew Cowie， Michael Figurnov， Fabian B. Fuchs， Hannah Gladman， Rishub Jain， Yousuf A.Khan， Caroline M. R. Low， PerlinKuba， PotapenkoAnna， SavyPascal， SinghSukhdeep， SteculaAdrian， ThillaisundaramAshok， TongCatherine， YakneenSergei， D. ZhongEllen， ZielinskiMichal， ŽídekAugustin， BapstVictor， KohliPushmeet， JaderbergMax， Demis Hassabis and John M.Jumper.2024 Accurate structure prediction of biomolecular interactions with AlphaFold 3. Nature， 630， 493–500 . ［https：//doi.org/10.1038/s41586-024-07487-whttps://doi.org/10.1038/s41586-024-07487-w］
Lake B. M. ， Salakhutdinov R. and Tenenbaum J. B. 2015. Human-level concept learning through probabilistic program induction. Science. 350（6266）：1332-1338. ［DOI：10.1126/science.aab3050http://dx.doi.org/10.1126/science.aab3050.］
Lechner M.， Hasani R.， Amini A.， Henzinger T.A.， Rus D.， Grosu R.. 2020.Neural Circuit Policies Enabling Auditable Autonomy. Nature Machine Intelligence. 2， 642–652 . ［DOI：https：//doi.org/10.1038/s42256-020-00237-3http://dx.doi.org/https：//doi.org/10.1038/s42256-020-00237-3］
Li Deyi. 2005. Uncertainty artificial intelligence. 2nd ed. Beijing： National Defense Industry Press（ISBN：9787118039214）
李德毅. 2005.不确定性人工智能. 第1版. 北京： 国防工业出版社
Li Deyi. 2014. Uncertainty artificial intelligence. 2nd ed. Beijing： National Defense Industry Press（ISBN：9787118090819）
李德毅. 2014.不确定性人工智能. 第2版. 北京： 国防工业出版社
Li Deyi. 2021.Ten questions and answers for the new generation of artificial intelligences［J］. CAAI Transactions on Intelligent Systemsx. 16（5）： 828-833. ［DOI： 10.11992/tis. 202103044http://dx.doi.org/10.11992/tis.202103044.］
Li Deyi. 2023a. Cognitive physics： the enlightenment by schrödinger， turing， and wiener and beyond . Intelligent computing.2： 0009.
Li Deyi. 2022. Creating a warm agricultural machinery with AI. CAAI Transactions on Intelligent Systems，17（05）：859.
李德毅.2022.用人工智能打造有温度的农机.智能系统学报，17（05）：859.
Li Deyi， Gan Wenyan， Liu Luying. 2003. Artificial Intelligence and Cognitive Physics . The 10th National Academic Annual Conference of the Chinese Association of Artificial Intelligence. Guanzhou，China： 6-15.
李德毅， 淦文燕， 刘璐莹.2003.人工智能与认知物理学［C］//中国人工智能学会第10届全国学术年会. 广州，中国：6-15
Li Deyi， Gao Hongbo. 2018. A Hardware Platform Framework for an Intelligent Vehicle Based on a Driving Brain.Engineering，4（04）：78-92.
李德毅，高洪波.2018.基于驾驶脑的智能驾驶车辆硬件平台架构.Engineering，4（04）：78-92.
Li Deyi， He Wen. 2023b. The troubles and release of intelligence Chinese Science. SCIENCE CHINA Information Sciences.
李德毅， 何雯. 2023. 智能的困扰和释放. 中国科学： 信息科学［DOI：10.1360/SSI-2023-0043http://dx.doi.org/10.1360/SSI-2023-0043.］
Li Deyi， Liu Yuchao， Ren Lu. 2023c. Artificial Intelligence Looking at Intelligence . Science and Society.13 （4）： 131-149（李德毅，刘玉超，任璐.2023人工智能看智慧.科学与社会， 13（4）：131-149.
Li Deyi， Meng Haijun， Shi Xuemei. 1995. Subordinate Cloud and Subordinate Cloud Generator . Computer Research and Development， 32 6
） （李德毅，孟海军，史雪梅.1995.隶属云和隶属云发生器.计算机研究与发展，32（6）
Li Deyi， Yu Quan， Jiang Guangjie. 1997. Unified Evaluation of Reliability， Durability， and Interference Resistance of C-3I Systems. Systems Engineering Theory and Practice， 17（03）：24-28
李德毅，于全，江光杰.1997.C~3I系统可靠性、抗毁性和抗干扰的统一评测.系统工程理论与实践，17（03）：24-28
Li Deyi， Yin Jialun， Zhang Tianlei， Han Wei， and Bao Hong. 2023d. The Four Elements Theory of Machine Cognition. Basic Science of China， 25（3）： 1-10，22
李德毅，殷嘉伦，张天雷，韩威，鲍泓.2023.机器认知四要素说［J］.中国基础科学， 25（3）：1-10，22.
Li Deyi， Zheng Siyi， Huang Liwei and Liu Yuchao. 2024. Cognition Formalization. China Basic Science，26（02）：1-14+28.
李德毅，郑思仪，黄立威，刘玉超.2024. 认知的形式化.中国基础科学， 26（02）：1-14+28.
Liang T.， Pan W.， Bao H.， Fan X.， Li H. 2023. Bird’s Eye View Semantic Segmentation based on Improved Transformer for Automatic Annotation， KSII Trans. Internet Inf. Syst.17（8）.1996-2015 ［doi：10.3837/tiis.2023.08.002.］.
Lin Longnian. Discovering the cellular fabric of the brain location system ［J］. Science， 2015， 67 （1）： 5. （林龙年.2015. 发现大脑定位系统的细胞组构.科学， 67（1）：5 ［DOI： 10.3969/j.issn.0368-6396.2015.01.009http://dx.doi.org/10.3969/j.issn.0368-6396.2015.01.009］
Liu K.‡， Zheng Y.‡， Yang J.Y.， Bao H.*， Zeng H.M.. 2021. Chinese Traffic Police Gesture Recognition Based on Graph Convolutional Network in Natural Scene. Applied Sciences. 11（24） ： 11951-11951.
相关作者
相关机构
京公网安备11010802024621