Brain-Inspired Systems: A Transdisciplinary Exploration on Cognitive Cybernetics, Humanity, and Systems Science Toward Autonomous Artificial Intelligence

This paper presents a panel summary on the framework of Autonomous Systems (AS) and paradigms in development. AS are advanced intelligent systems and general AI technologies triggered by the transdisciplinary development in intelligence science, system science, brain science, cognitive science, robotics, computational intelligence, and intelligent mathematics. It is recognized that, in a rigorous perspective, the only matured AS is human brains and human collective intelligence. It explains why there was rarely man-made AS in the past half century, because of the theoretical, mathematical, computational, and programming language unreadiness. Therefore, the ultimate goal of AS is to implement a brain-inspired system that may think and behave as a human counterpart in hybrid intelligent systems and general AI implementations. There is no doubt that AS will be increasingly demanded by the intelligence-based industries and societies for cognitive computers, deep machine learning systems, robotics, brain-inspired systems, mission-critical systems, self-driving vehicles, and intelligent appliances.

Brain-Inspired Systems (BIS) are an emerging field of brain and intelligence sciences that studies natural intelligence models of AI and cognitive systems in one direction, and the formal models of the brain simulated by computational intelligence in another direction. A typical BIS is the cognitive robots that mimic and implement the brain through all cognitive levels. BIS provides insights for brain-machine interfaces (BMI), which may lead to novel man-machine interactions and hybrid intelligent systems. BIS may also advance classic computers from dada processors to the next generation of knowledge processors mimicking the brain. BIS will underpin a wide range of engineering paradigms such as cognitive systems, cognitive computers, cognitive robots, machine learning systems, semantic comprehension systems, big data systems, unmanned systems, self-driving vehicles and hybrid man-machine systems.

This presentation concerns some idea of what could be done, in the author's view, to help make Wang's cognitive informatics a powerful and viable source of tools and techniques for solving various real life problems. First, we give a brief account of cognitive informatics meant as a multidisciplinary field within informatics, or computer science, that is based on results of cognitive and information sciences, and which deals with human information processing mechanisms and processes and their decision theoretic, engineering, etc. applications in broadly perceived computing. We focus on its purpose, i.e. to develop and implement technologies to facilitate and extend the information acquisition, comprehension and processing capacity of humans. Emphasis is on underlying processes in the brain. However, we advocate an extended approach in which though the very cognitive informatics is the foundation, as those processes in the brain are crucial, some sort of an “outer” cognitive informatics is needed which explicitly makes reference not what proceeds “internally” in the brain, because we do not “see” this, but “externally”, i.e. what people can see, judge, evaluate, etc., and what is clearly a result of cognitive information specific processes in the brain. This line of reasoning is in line with the very essence of comprehension, memorizing, learning, choice and decision making, satisfaction with partial truth, allowing for not perfect solutions, etc. dealt with using tools and techniques derived from many areas like psychology, behavioral science, neuroscience, artificial intelligence, linguistics, neuroeconomics etc. In our case, we will concentrate on some cognitive informatics type elements that mostly have been inspired by psychology and behavioral sciences, as our problem is inherently related to human judgments and perceptions, but we will mentioned some inspirations from neuroscience, notably along the lines of neuroeconomics. Cognitive informatics constitutes a foundation of its related new field, cognitive computing, which is basically a new direction in broadly perceived intelligent computing and systems that synergistically combines results from many areas, e.g., information science, computational sciences, computer science, artificial and computational intelligence, cybernetics, systems science, cognitive science, (neuro)psychology, brain science, linguistics, etc. to just mention a few. We try to show on an example of a dynamic systems modeling, more specifically scenario based regional development planning, that cognitive computing can provide new conceptual and implementation vistas. Basically, we consider a region that is characterized by 7 life quality indicators related to economic, social, environmental, etc. qualities, which evolve over some planning horizon due to some investments, mostly by some regional or governmental agencies. There are some scenarios of investment levels over the planning horizon, meant for the development of the particular life quality indexes, and some desired levels of these indexes, both objective, i.e. set by authorities, and subjective, i.e. perceived by the inhabitant groups. As a result of a particular investment scenario, the life quality indexes evolve over the planning horizon, and their temporal evolution is evaluated by the authorities and inhabitants. This evaluation has both an objective, i.e. against the “officially” set thresholds, and subjective, i.e. as perceived by various humans and their groups. Basically, we employ Kacprzyk's fuzzy dynamic programming based approach to the modeling and planning/programming of sustainable regional development, with soft constraints and goals, but we advocated a more sophisticated assessment of variability, stability, balancedness of consecutive investments. In this process we try to develop evaluation measures, and then the optimization type model using concepts that can be effectively and efficiently handled by cognitive computing, notably the inclusion of the so called decision making and behavioral biases, biases in probability and belief, social biases, memory errors, etc. Moreover, we strongly reflect the so called status quo and minimal change biases. By using many results from social sciences, psychology, behavioral economics, neuroeconomics, etc. on human judgments and human centric evaluations, we augment a traditional purely effectiveness and efficiency oriented analysis by a more sophisticated analysis of effects of variability of temporal evolution of some life quality indicators on the human perception of its goodness. The model presented, which has been employed for years as part of large mathematical modeling projects for sustainable regional development in many regions in Asia and Europe, is illustrated on an example with scenario analysis for a rural region plagued by social and economic difficulties in which subsidies should properly be distributed over time to obtain a best overall socioeconomic effect. In this talk we present the model in a different perspective, based first on the basic Wang's cognitive informatics and its Wang and Ruhe's decision making application, and then based on new, more comprehensive cognitive computing. We show that this provides a novel insight.

The Covid-19 pandemic reminds us again about our limited knowledge and understanding in the nature including both micro and macro worlds. We have been developing a variety of tools such as automation, robotics, internet, and artificial intelligence (AI), etc. to augment human capability for improved safety, quality, and productivity in work and life, but human lives are still vulnerable over 100 years since the last Spanish Flu in 1918. We are even more vulnerable when the tools we developed (e.g., automation and AI) do not understand human intent or follow human instructions. Recent accidents to the Boeing 737 Max passengers ring the alarm again about the imperative needs of appropriate design concepts and scientific methodologies for developing safety critical cognitive and/or autonomous systems or AI functions and collaborative partnership of human and intelligent systems. With AI and its related technologies reach their bottleneck, it is even more vital to follow scientific and systematic methodology to understand well about capacity and limitation of both human intelligence and machine intelligence so that their strengths can be optimized for a collaborative partnership when dealing with safety critical situations. This talk discusses about the needs for the researchers, designers, developers, and all practitioners who are interested in building and using 21st century human-autonomy symbiosis technologies (Why). It touches the topics of proper analytical methodologies for functional requirements of the intelligent systems, design methodologies, implementation strategies, evaluation approaches, and trusted relationships (How). These aspects will be explained with real-world examples when considering contextual constraints of technology, human capability and limitations, and functionalities that AI and autonomous systems should achieve (When). Audience will gain insights of context-based and interaction-centered design approach for developing a safe, trusted, and collaborative partnership between human and technology by optimizing the interaction between human intelligence and AI. The challenges and potential issues will also be discussed for guiding future research and development activities when augmenting human capabilities with AI, and cognitive and/or autonomous systems.

Message from Editors-in-Chief

In this paper, methodologies of building synergetic learning systems (SLS), which refer to artificial intelligence (AI) systems, are presented. From our viewpoint, an AI system might not only be constructed with deep learning but also with multi-discipline knowledge. To build the SLS, we need to develop the methodologies which should draw on the mathematical methods of neurocognitive mechanisms and machine learning, knowledge across a wide range of disciplines including cognitive neuroscience, physics, psychology, medicine, automation, computer science, life science, systems science and social science, and so on. It is dependent on the integration of statistical physics and the systematic view of complexity science. Our methodologies of building SLS are investigated systematically and analyzed on multi-scale, multi-level and multi-perspective. With our proposed methodologies, an AI system can be constructed with the properties of interpretability, extensibility and evolvability.

Autonomous systems are advanced intelligent systems and general AI technologies triggered by the transdisciplinary development in intelligence science, system science, brain science, cognitive science, robotics, computational intelligence, and intelligent mathematics. AS are driven by the increasing demands in the modern industries of cognitive computers, deep machine learning, robotics, brain-inspired systems, self-driving cars, internet of things, and intelligent appliances. This paper presents a perspective on the framework of autonomous systems and their theoretical foundations. A wide range of application paradigms of autonomous systems are explored.

Symbiotic autonomous systems (SAS) are advanced intelligent and cognitive systems that exhibit autonomous collective intelligence enabled by coherent symbiosis of human-machine interactions in hybrid societies. Basic research in the emerging field of SAS has triggered advanced general-AI technologies that either function without human intervention or synergize humans and intelligent machines in coherent cognitive systems. This work presents a theoretical framework of SAS underpinned by the latest advances in intelligence, cognition, computer, and system sciences. SAS are characterized by the composition of autonomous and symbiotic systems that adopt bio-brain-social-inspired and heterogeneously synergized structures and autonomous behaviours. This paper explores the cognitive and mathematical foundations of SAS. The challenges to seamless human-machine interactions in a hybrid environment are addressed. SAS-based collective intelligence is explored in order to augment human capability by autonomous machine intelligence towards the next generation of general AI, cognitive computers, and trustworthy mission-critical intelligent systems. Emerging paradigms and engineering applications of SAS are elaborated via autonomous knowledge learning systems that symbiotically work between humans and cognitive robots. This article is part of the theme issue 'Towards symbiotic autonomous systems'.

The Mindset of Cognitive Science.

Modelling our selves: Being No One: The Self-Model Theory of Subjectivity by Thomas Metzinger, MIT Press, 2003. $55.00/£34.95 (hbk) (634 pages) ISBN 0 262 13417 9

A cognitive system is any real system that has some cognitive property. Therefore, cognitive systems are a special type of K-systems (see chapter 3, section 3). Note that this definition includes both natural systems such as humans and other animals, and artificial devices such as robots, implementations of AI (artificial intelligence) programs, some implementations of neural networks, etc. Focusing on what all cognitive systems have in common, we can state a very general but nonetheless interesting thesis: All cognitive systems are dynamical systems. Section 2 explains what this thesis means and why it is (relatively) uncontroversial. It will become clear that this thesis is a basic methodological assumption that underlies practically all current research in cognitive science. The goal of section 3 is to contrast two styles of scientific explanation of cognition: computational and dynamical. Computational explanations are characterized by the use of concepts drawn from computation theory, while dynamical explanations employ the conceptual apparatus of dynamical systems theory. Further, I will suggest that all scientific explanations of cognition might end up sharing the same dynamical style, for dynamical systems theory may well turn out to be useful in the study of all types of models currently employed in cognitive science. In particular, a dynamical viewpoint might even benefit those scientific explanations of cognition which are based on symbolic models. Computational explanations of cognition, by contrast, can only be based on symbolic models or, more generally, on any other type of computational model. In particular, those scientific explanations of cognition which are based on an important class of connectionist models cannot be computational, for this class of models falls beyond the scope of computation theory. Arguing for this negative conclusion requires the formal explication of the concept of a computational system that I gave in chapter 1 (see definition 3). Finally, section 4 explores the possibility that scientific explanations of cognition might be based on Galilean models of cognitive systems (see chapter 3, section 5). Most cognitive scientists have not yet considered this possibility. The goals of this section are to contrast this proposal with the current modeling practice in cognitive science, to make clear its potential benefits, and to indicate possible ways to implement it.

With ever increasing complexity of products and customer demands, companies are adopting new strategies to meet the changing technological requirements, shorter product life cycles, and globalization of manufacturing operations. Product design and development are getting more sophisticated procedures/processes involved and require designers and engineers possessing different expertise, knowledge and experience to work together [1,2]. To address these challenges, techniques based on artificial intelligence (AI) are increasingly being used to improve effectiveness and efficiency in the product design and development life cycle. Intelligent systems can be beneficially applied to many aspects of design and also design-related tasks, for example, identifying customer demands and requirements, design and planning, production, delivery, marketing and customer services, etc. Individual intelligent paradigms (such as fuzzy logic, neural network, genetic algorithm, case-based reasoning, and especially expert systems) have been applied to specific stages of the design process (especially expert systems). However, increasingly, hybrid solutions, that integrate multiple individual intelligent techniques, are required to solve complex design problems. The integrated and hybrid intelligent environment can provide various information and knowledge for supporting rapid and intelligent decision-making throughout the entire design and development process. This is in line with the evolutionary trends of product design and development process, from the traditional CAD systems into the knowledgebased engineering and integrated intelligent design systems through a combination of concurrent engineering, collaborative engineering and integrated and hybrid AI techniques. In recent years, with the advancement of AI and intelligent systems techniques, integrated and hybrid intelligent systems are gaining better acceptance in product design and development. The driving force behind this is that integrated and hybrid intelligence and distributed 3C (collaboration, cooperation, and coordination) allow the capture human knowledge and the application of it so as to achieve high quality designs/products. Further motivation arises from steady advances in individual and hybrid intelligent-systems techniques, and the widespread availability of computing resources and communications capability through intranets and the World Wide Web [1]. This special issue aims to collect relevant original works in the application of emerging integrated and hybrid intelligent systems techniques to the design of products, processes and systems in product development. The goal is to take a snapshot of the progress in the research into the support for product design and development and to disseminate how recent developments in integrated, knowledge-intensive and computational AI techniques can improve and enhance such support. The selected papers provide an integrated, holistic perspective on this complex set of challenges and provide rigorous research results. The focus of this special issue is on the integrated and hybrid intelligent methodologies, frameworks and systems for support-

Submit an article entitled thus to your publishing editor and they will raise the query: Give the eponymous sum to a schoolchild and, provided that he/she has learned to add (or has learned to use a calculator), he or she will also tell you the answer is 105721. Alan Turing was unabashed, in a 1950 article1 in Mind, to put the incorrect answer into the mouth of a fictional player of the ‘imitation game’. The article was entitled ‘Computing, machinery and intelligence’. In it Turing asked ‘Can a machine think?’ The game came to be known as the Turing test, through which a congruence of machine and human intelligence could be established. Turing's suggested route to intelligent machines was to mimic the machinery of a child's brain, and educate the resulting machine. That suggestion continues to inspire researchers – and robots.2 And now there's a new kid on the block! WolframAlpha is described – carefully, without mentioning intelligence – as a ‘computational knowledge engine’. Its goal is to3 build on the achievements of science and other systematizations of knowledge to provide a single source that can be relied on by everyone for definitive answers to factual queries. WolframAlpha is said to contain more than 10 trillion pieces of data, more than 50,000 algorithms and models, and linguistic capabilities from more than 1000 domains. It's great fun to play with. One commentator has suggested that WolframAlpha is possibly an ‘emerging artificial intelligence and a step towards a self-organising internet’. That's most unlikely, at least if we use the Turing test as the criterion for judging the emergent artificial intelligence. But it's only a failure if the aim is that AI should compete with HI – human intelligence. It appears very much the mission of this journal that the aim is to complement and complete HI. It is with great pleasure that we welcome to the board two new members: Professors Adrian Hopgood (De Montfort University, UK) and John Fox (University of Oxford and University College London, UK). Professor Adrian Hopgood is Dean of the Faculty of Technology at De Monfort University, Leicester, UK. His research interest concerns artificial intelligence, including knowledge-based and distributed multi-agent systems, computational intelligence (artificial neural networks, genetic algorithms, fuzzy logic) and their practical applications. He has a particular passion for hybrids, which bring together the best techniques for particular applications. As well as a PhD, Adrian has a Diploma in French and an MBA from the Open University. He is also author of the best selling Intelligent Systems for Engineers and Scientists. Professor John Fox attended Durham (UK) and Cambridge Universities (UK) and held postdoctoral fellowships at Carnegie-Mellon and Cornell Universities in the USA. After returning to the UK in 1975 he worked on decision making and artificial intelligence in medicine, joining Cancer Research UK (then ICRF) in 1981 to set up an interdisciplinary group in artificial intelligence, computer science and medicine. John's interests include cognitive science, computing and biomedical engineering. A recent book Safe and Sound: Artificial Intelligence in Hazardous Applications deals with the use of artificial intelligence in medicine and other safety-critical fields. He was founding editor of the Knowledge Engineering Review. John is now at Oxford where he has set up a new collaboration in cognitive science and systems engineering (http://www.cossac.org). Please welcome both Adrian and John to the board. This month we have four excellent papers. In ‘Fuzzy based fast dynamic programming solution of unit commitment with ramp constraints’, Patra et al. give insight into the use of fuzzy logic in the solution to difficult-to-solve multi-stage decision-making problems, showing that fuzzy models can perform as well on these problems as more traditional techniques. The difficulty of obtaining realistic simulated real-world data suitable for the validation of models used in knowledge engineering of medical solutions is well known. In ‘Preliminary evaluation of electroencephalographic entrainment using thalamocortical modelling’, Cvetkovic et al. consider a number of theoretical models for their ability to replace actual data. Importantly, the authors identify a number of shortcomings that should be overcome for advances in this area of knowledge engineering to be made. In ‘Modified mixture of experts employing eigenvector methods and Lyapunov exponents for analysis of electroencephalogram signals’, Übeyli presents another in her excellent series of papers dedicated to detecting variability of electroencephalogram signals, this time with a classification accuracy of 98.33%. Continuing the theme of the second paper, if not the approach, Yang and Kecman consider the use of an adaptive local hyperplane algorithm as a suitable classifier when only a small data set is available to learn from. The proposed classifier outperforms, on average, all the other four benchmarking classifiers in this area. Enjoy!

The hierarchy of human knowledge is categorized at the levels of data, information, knowledge, and intelligence. For instance, given an AND-gate with 1,000-input pins, it may be described very much differently at various levels of perceptions in the knowledge hierarchy. At the data level on the bottom, it represents a 2 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1,000</sup> state space, known as `big data' in recent terms, which appears to be a big issue in engineering. However, at the information level, it just represents 1,000 bit information that is equivalent to the numbers of inputs. Further, at the knowledge level, it expresses only two rules that if all inputs are one, the output is one; and if any input is zero, the output is zero. Ultimately, at the intelligence level, it is simply an instance of the logical model of an AND-gate with arbitrary inputs. This problem reveals that human intelligence and wisdom are an extremely efficient and a fast convergent induction mechanism for knowledge and wisdom elicitation and abstraction where data are merely factual materials and arbitrary instances in the almost infinite state space of the real world. Although data and information processing have been relatively well studied, the nature, theories, and suitable mathematics underpinning knowledge and intelligence are yet to be systematically studied in cognitive informatics and cognitive computing. This will leads to a new era of human intelligence revolution following the industrial, computational, and information revolutions. This is also in accordance with the driving force of the hierarchical human needs from low-level material requirements to high-level ones such as knowledge, wisdom, and intelligence. The trend to the emerging intelligent revolution is to meet the ultimate human needs. The basic approach to intelligent revolution is to invent and embody cognitive computers, cognitive robots, and cognitive systems that extend human memory capacity, learning ability, wisdom, and creativity. Via intelligence revolution, an interconnected cognitive intelligent Internet will enable ordinary people to access highly intelligent systems created based on the latest development of human knowledge and wisdom. Highly professional systems may help people to solve typical everyday problems. Towards these objectives, the latest advances in abstract intelligence and intelligence science investigated in cognitive informatics and cognitive computing are well positioned at the center of intelligence revolution. A wide range of applications of cognitive computers have been developing in ICIC [http://www.ucalgary.ca/icic/] such as, inter alia, cognitive computers, cognitive robots, cognitive learning engines, cognitive Internet, cognitive agents, cognitive search engines, cognitive translators, cognitive control systems, cognitive communications systems, and cognitive automobiles.

Use in clinical practice of an automated screening method of diabetic retinopathy that can be derived using a diagnostic artificial intelligence system