Artificial cognitive systems of the present and future

    Hello!

    As we reported earlier, Innopolis University organized a series of open webinars on various topics related to Computer Science.

    The first webinar on the topic of Artificial Intelligence took place on February 11, and was hosted by Associate Professor of the University of Samir Belhauari . We were pleasantly surprised by the activity of the participants asking interesting, non-trivial questions. A full webinar entry is at the end of this post. By the way, please note that the application campaign contest to study at undergraduate Innopolis University is in full swing, hurry up.

    Today we want to talk about our next webinar - it will be held on February 18 at 18:00 (Moscow time) and will be dedicatedArtificial cognitive systems (Artificial Cognitive Systems). Registration is here . It will be held by visiting professor at Innopolis University, David Vernon . No one will tell about himself better than the author himself, so we decided to post here a translation of the material from his personal website, in which Professor Vernon talks about the scope of his scientific activity. So, a word to our teacher:

    Although most of my work over the past 28 years has been related to computer vision and AI, starting with my dissertation on robotics and robot vision, I began to study cognitive and autonomous systems immediately after my defense of Ph.D. in 1985. The work of Maturana and Varela and their principles of autopoiesis and surgical closure had a significant impact on my thinking, as did the philosophy of Husserl and Heidegger phenomenology. Unfortunately, despite many attempts to develop and build an autopoiesis system, the next 20 years have yielded very little to support empirical scientific results. I published several scientific papers on this subject, but they did not satisfy me as an engineer, and did not impress the scientific community, which was still “captured” by the hypothesis of the physical systems of Newell and Simon, and its accompanying approach to cognition and artificial intelligence. Therefore, pragmatism prevailed in me (but only for a while), and I focused on computer vision. This activity brought the result in the form of a new Fourier-technologies for recognition and segmentation of fuzzy images.



    During this quarter-century period, however, I observed a slow but inevitable change in the thinking of the scientific community. In recent years, it has become apparent that the theory of knowledge is very intertwined with the physical structure of the body and its interaction with the external environment. Intelligence and thought processes are deeply dependent on body structure, motor functions, and especially manipulation skills, as well as muscle elasticity and morphology of the retina and sensory system. The human body and its actions play no less a role in cognition than neural processes, and human intelligence develops through interaction with objects of the environment and other people.

    This new look at AI is a fundamental shift in the transition from functionalism and the dualism of cognitivism and “classical” artificial intelligence to an alternative concept that once again emphasizes the superiority of materialism, development and interaction in cognitive systems. These models are mainly based on the theory of dynamical systems, complex forms of connectivism and self-organization. All of these concepts resonate strongly with the early writings of Maturana and Varela. In a generalized form, these concepts are known as enactive systems .

    Active systems are based on five central principles - materialism, experience, emergence, autonomy and understanding. Active approaches physically and physically play a key role in cognition. Consequently, it is impossible not to take ontogenesis into account as an objective fact, since the experience of the observer determines his cognitive understanding of the world in which he exists.

    One of the most tangible consequences of this change in philosophy is the implementation of interdisciplinary research, combining neuroscience, age psychology, epigenetic robotics, the theory of complex systems, philosophy, and, of course, computer science. The European Commission actively supports and stimulates the development of this new science. I can say that I am lucky to take part in two Commission projects - RobotCub and euCognition.



    RobotCub is creating a cognitive humanoid robot - iCub, the main goal of this project is to create an open platform and research in the field of active cognition. In this project, I am creating the cognitive architecture of this robot. I hope that finally, after 25 years, today we have the tools and embryonic theories that will allow us to implement a truly functioning, effective cognitive system that would pay tribute to the ideas of the pioneer of this science - Francisco Varela.



    Recording of the Artificial Intelligence webinar February 11, 2014 (duration - 39 minutes)

    Decoding of questions and answers during the webinar: vk.com/topic-56385969_29710758

    Also popular now: