Autonomy-Oriented Computing (AOC), Self-organized Computability, and Complex Data Mining

Future data-mining challenges will lie in the breakthroughs in new computational paradigms, models, and tools that can offer scalable and robust solutions to complex data-mining problems. The problems of such a nature can be: (1) petabyte-scale (e.g., mining Google books or social networks), (2) dynamically-evolving (e.g., detecting air traffic patterns, market trends, and social norms), (3) interaction-rich as well as trans-disciplinary (e.g., predicting and preventing world economic and/or ecological crisis), and/or (4) highly-distributed (e.g., security and adaptive computing, such as community evolution, in pervasive environments). Toward this end, various computing ideas and techniques have been proposed and explored that explicitly utilize the models of computational autonomy as inspired by nature. This talk focuses on one of such research initiatives, which concerns the development of an unconventional computing paradigm, called Autonomy-Oriented Computing (AOC). In general, AOC tackles a complex computing problem by defining and deploying a system of local autonomy-oriented entities. The entities spontaneously interact with their environments and operate based on their behavioral rules. They self-organize their structural relationships as well as behavioral dynamics, with respect to some specific forms of interactions and control settings. Such a capability is referred to as the self-organized computability of autonomous entities. In this talk, we will examine basic concepts and principles in the development of an AOC system, and present some related data-mining examples in the area of complex networks, e.g., unveiling community structures, characterizing the empirical laws of WWW user behavior, and/or understanding the dynamic performance of social networks.