My current mission is to understand the emergence and the organizational principles of complex adaptive systems, both natural and artificial. While being interested in a wide range of phenomena in nature and society, my focus is on the study of the human brain and cognition. My mission is motivated and driven by the following questions: What are the principles that would enable us to understand how the human brain models reality and self? How neuromolecular and neurophysiological processes give rise to psychological phenomena and how the physical and molecular/biological bases of these phenomena can be adequately modelled and further integrated with psychological theories? To what extent and under what conditions can we implement our knowledge about neurocognition into in-silico artificial organisms and virtually embodied autonomous agents and thereby test our neurocognitive integrative models?

More specifically, my current research deals with 1) empirical studies of human information processing, 2) mathematical/neural models of spatiotemporal aspects of visual, linguistic and memory processes and learning, 3) agent-based modeling and simulation, 4) evolution of cooperation and evolutionary game theory, and 5) evolution of language. I started my research with visual sensory memory experiments, in which I investigated the effects of physical features of visual and auditory stimuli on recall performance in linguistically different populations. These experiments also lead me to the conclusion that the human loss of (visual sensory and working) memory capacity in the course of evolution was primarily due to the emergence of other memory-related skills such as representation and hierarchical organisation, and most notably, language. These results were also more recently confirmed (Inoue, S. & Matsuzawa, T., Current Biology, 17 (2007) 1004-1005).

In the area of agent-based modeling (ABM), I am investigating the physical and behavioral bases of competition, conflict emergence and cooperation in social groups. ABM is a computational model for simulating the actions and interactions of autonomous individual agents in a network, in an attempt to re-create and predict the complex behavioral patterns of an investigated system at a more global level. It combines methods and ideas from game theory, complex systems, emergence, computational sociology, sociobiology, statistical physics, multi agent systems, and evolutionary programming. My recent focus in this field has been on evolution of cooperation on large-scale complex networks, particularly in the context of in-group altruism (by phenotypic similarity). This work on ABM is done jointly with Prof. Dr. Dr. h.c. Dietrich Stauffer and Prof. Dr. Christian Schulze. Professor Stauffer holds the world record in the size of simulated Ising models (he simulated 1000102^2, 9984^3, 880^4, 176^5 and 48^6 Ising model, and 1000192^2 cellular automata (speed: 145 GUPS; 550 GUPS for 640*640). For percolation, Daniel Tiggemann (PhD thesis in 2006 under the supervision of Prof. Stauffer) reached (7 million)^2, 25024^3, 1305^4, 225^5; and (2 million)^2 for Ising).

With Klaus Lichtenegger (Dept. of Theoretical Physics, University of Graz, Austria), I am extending some previous work (Hadzibeganovic, Stauffer, & Schulze, 2008) on ABM in the context of Hopfield neural networks with up to 10^8 nodes. Besides classical applications to opinion dynamics (e.g. opinion flow modeling in n-party systems), in this collaboration we aim to provide simulations of a number of human features enabled via the activity of mirror neurons in the human brain, including imitation, empathy, mindreading, and language learning.

More recently, together with Dr. Gudberg K. Jonsson and Prof.Dr. Magnus S. Magnusson (Human Behavior Laboratory, University of Iceland), I am using a new pattern detection and analysis system for measuring and comparing the emerging behavioral patterns across human and artificial populations in soccer games. We intend to standardize this pattern detection system as a measure of efficiency of research improvements that could be used by all research groups within the RoboCup community.

Under the supervision of Prof.Dr. Dietrich Albert (Cognitive Science Section, Dept. of Psychology, University of Graz, Austria), I am working on eye-tracking studies of language processing and learning. The empirically obtained eye-tracking data are modeled by using neural networks or other approaches. Also, in the area of language research, I've been collaborating with Dr. Aida Vidan (Harvard University, USA) and Prof.Dr. Victor Friedman (University of Chicago, USA), on projects related to (quantitative) linguistic typology, language contact and conflict.

In the area of artificial language evolution in autonomous robots, I’ve been investigating various aspects of language emergence and evolution by considering an embodied cognition approach to this problem. Initially, my focus here was on the spatio-temporal aspects of conceptual development in robotic agents, but recently I turned more to studying the emergence of categories and coordination of linguistic inventions in situated interactions. My particular focus has been on creating more realistic agent-based language and categorization game experiments simulated on large-scale complex networks.

Most recently, Prof.Dr. Taiki Takahashi (Dept. Behavioral Science, University of Hokkaido, Japan) and I have started researching on further applications of the generalized entropy of Tsallis, particularly in the areas of cognitive science and neuroeconomics. The focus of this research collaboration is on cultural differences in temporal and probability discounting, but also on foundational issues in neuroeconomics.

I am also researching in the areas of nonextensive neural network modeling of (language) learning (work done jointly with Prof.Dr. Sergio A. Cannas) and biologically-inspired quantum neural-like network systems for visual information processing (work done jointly with Dr. Chu Kiong Loo).

In my studies of visual object processing, I use Quantum Neural-like Networks (QNNs). Here, I model the viewpoint-invariant appearance-based object processing along the human visual pathway (retina, lateral geniculate nucleus, primary visual cortex) up to the inferior occipito-temporal cortex. QNNs represent a new and a very fascinating approach to understanding cognition based on quantum information processing principles. QNNs were inspired by the Holonomic Brain Theory of Prof. Dr. Dr. h.c. mult. Karl H. Pribram, the founder of experimental neuropsychology and one of the founding fathers of modern neuroscience.

Nonextensive neural network models for language learning employ the nonextensive statistics theory of Prof. Dr. Dr. h.c. Constantino Tsallis. In my investigations of human and artificial learning, I use this kind of statistics to include some degree of nonlocality in simple, two-level neural nets. The resulting behavior of the model raises the possibility of using entropic nonextensivity as a means of characterizing the degree of complexity of learning in both natural and artificial systems. The model may further be useful in developing diagnostic monitoring tools that could be applied in a variety of learning domains, but also in studying other problems in neuroscience such as neurological impairments.

More recently, my work has been cited in Ben Goertzel's book The Hidden Pattern (2006, Brown Walker Press), in Constantino Tsallis’ book Introduction to Nonextensive Statistical Mechanics: Approaching a Complex World (2009, Springer Verlag), in Lecture Notes in Computer Science (Volume 4681 (2007) 25-33), in Physica A: Statistical Mechanics and its Applications (Volume 387 (2008) 3242–3252), in Journal of Image and Graphics (Volume 13 (2008) 119-123), and again in Physica A: Statistical Mechanics and its Applications (Volume 388 (2009) 174–186). My Erdös number is 5, my Mandelbrot number is 2. My long-term goal is to implement the major results of my studies in the area of advanced and adaptive artificial intelligence.