Methods of complexity reduction and their effects in social, economic, and computational environments

Since the world around us is too complex to understand and handle every aspect and detail, we need to reduce its complexity and simplify. Often, this is achieved by structural assumptions, and in interactions, these structural assumptions may in turn shape the interactions. Concrete examples to be discussed in this talk include word meanings in language, or interactions in financial markets (we have conducted interviews with investment and central bankers to find out how they cope with strategic uncertainty in situations of reciprocal bounded rationality). But suitable structural assumptions also occur as gestalt rules in psychology, and can constitute the key to the success of various machine learning schemes. Perhaps, eventually a general theory can be achieved.