Many robotic systems are derived from and can be imitated by social structures of beings. Examples of this are transportation belts, light cranes, multiple beings instead of ladders/scaffoldings. Even ants can build bridges using their bodies. Post-earthquake works are a critical application, where heavy machinery cannot be used among unstable ruins, or blocked streets. Instead of letting slow and inefficient human rescuers to carry out works, social structures of rescue-robots can be proposed. Such rescue-robots (supervised by a trained operator) can configure on demand into fast working equivalents of excavators, light cranes (to remove and lift debris) scaffoldings to extract victims from upper floors or e.g. to stabilize a crumbling wall, etc. Moreover, rescue-robots can deliver heavy detectors used to find victims, as well as heavy duty e.g. jaws-of-life to cut concrete and steel. The question now is how to define an individual and Collective Intelligence of such rescue-robots as well as their behavioral patterns, and how such robots should be shaped. The answer to the first part of this question is provided by the computational model of Adam Smith’s Invisible Hand phenomenon. An altogether new type of Collective Intelligence algorithms emerges on this basis. ACO algorithms are only a harbinger of a domain of such algorithms. The talk will present how such algorithms and social structures of robots should be designed and how they can be modeled to provide the ability to verify different possible scenarios of rescue on post-earthquake site.
Tadeusz (Tad) Szuba is engaged in research on machine intelligence since his Ph. D. dissertation in 1979. Around 1982 he started to work on an intelligent, quasi autonomous robot-excavator. To overcome technical and budget difficulties with providing prototypes of such robots, he started to deal with animated 3D graphics, where in Virtual Space such robots will work. He quickly became a renowned specialist in animated 3D graphics and Virtual Reality. Between 1992-1994 he set up a computer graphics laboratory using powerful Silicon Graphics workstations for Kuwait University, to teach students and to simulate a robot to find, post Gulf-War landmines and to destroy them. Between 1992-2001 he was working on a computational model of Collective Intelligence, which was concluded with the monograph: Szuba T.: Computational Collective Intelligence, Wiley & Sons NY, 2001. Since 2001 he has been working on a computational model of Adam Smith’s Invisible Hand phenomena. Since this problem is highly conflicting in area of economy and social science (advocates of liberalism vs advocates of interventionism) he has found a safe application haven in area of social structures of rescue-robots to support rescue actions after earthquake like in Amatrice, Italy, 2016. His hobby is skiing, tennis and sailing.
Network-Oriented Modeling is a relatively new way of modeling that is especially useful to model intensively interconnected and interactive processes. It has successfully been applied to model networks for a wide range of phenomena, including biological networks, networks of mental states, and social networks. In this lecture this modeling perspective will be discussed in more detail. It is discussed how the interpretation of a network as a causal network and taking into account dynamics brings more depth in the perspective. In the obtained notion of a temporal-causal network, nodes represent states with values that vary over time, and connections represent causal relations describing how states affect each other. As these causal relations themselves also may change, adaptive networks are covered as well. The wide scope of applicability of such a Network-Oriented Modelling approach will be analyzed in more depth and illustrated. This covers, for example, network models for principles of social contagion or information diffusion, and adaptive network models for principles of Hebbian learning in networks of mental states but also for principles of evolving social networks, such as the homophily principle and the triadic closure principle. From the methodological side, it will be discussed how mathematical analysis can be used to identify the relation between emergent dynamic properties concerning stabilizing or limit behaviour and network structure and settings. Finally, it will be discussed how requirements specification and verification can play an important role in the design process of a network model.
Since 1986 Jan Treur works in Artificial Intelligence, from 1990 as a full professor at the VU University Amsterdam; he is leading the Behavioural Informatics Group within the Artificial Intelligence Section. He is an internationally well-recognized expert in human-directed AI and cognitive and social modelling. His research during the past 10 years concerns both fundamental and application-directed aspects of human-directed AI. This covers methods and techniques for modelling and analysis of human-directed AI systems in a number of application areas, including Cognitive and Social modelling and simulation. Currently his research includes Network-Oriented Modeling approaches based on temporal-causal networks to model cognitive, affective and social interactions, with a book about this published in 2016. Modeling, analysis and simulation addresses domains in other scientific disciplines such as Biology, Neuroscience, Cognitive Science, and Social Sciences. Applications cover human-aware or socially aware AI systems and virtual agents.