miller@santafe.edu
Background
Some Projects
|
|
Ex Machina: Coevolving Machines and the Origins of the Social Universe |
|
|
Complex Adaptive Social Systems |
|
|
Discovering Novel Chemotherapies |
|
|
Giving According to GARP |
Computational Modeling
Selected Research Documents
| Working Papers | |||||||||
|
|
Books | ||||||||
|
|
Published Papers
|
Institutional Affiliations:
Carnegie Mellon
SDS
Santa Fe Institute
Background
My research focuses on understanding the behavior that emerges in complex adaptive social systems (CASS). Understanding the behavior of CASS, composed of interacting, thoughtful (but perhaps not brilliant) agents, is one of the great challenges of science. Such systems capture pervasive and important phenomena, arising in biological, chemical, environmental, economic, organizational, and political systems. Traditional scientific approaches to understanding these types of systems have shown only a limited ability to pry away their secrets. Part of this failure is attributable to a fundamental limitation of the reductionist approach to science: reducing systems to components does not imply the ability to reconstruct them (as anyone who has scrambled an egg innately understands). While it may seem that because we understand one we must also understand two, because one and one make two, this is only true if we also understand and.
Colleagues and I have applied CASS models to the analysis of a variety of key social phenomena. For example, we have explored the dynamics of political platforms in spatial elections, how decentralized institutional mechanisms can be used to sort agents into coherent groups, the emergence of cooperation in the iterated Prisoner's Dilemma, strategic choice in simple two-person games, the development of strategic communication, and bidding behavior in auction markets. To understand the fundamental dynamics of price formation in simple markets, we designed and organized an international computerized double auction tournament in 1988-90 (which was also the first internet-based auction market---here is the DAT Participant's Manual). This tournament allowed us to create an "artificial world" of trading agents in which we explored a variety of theoretical and practical issues. All of the above work indicates the remarkably potential for CASS modeling to advance our understanding about some of the most fundamental questions in the social sciences.
For CASS modeling, traditional tools of scientific theory building, such as mathematical modeling, need to be supplemented with computational models that take advantage of recent technological progress. As the mathematician Stanislaw Ulam pointed out, ``[T]he use of computers seems thus not merely convenient, but absolutely essential...I believe that the experience gained as a result of following the behavior of such processes will have a fundamental influence on whatever may ultimately generalize or perhaps even replace in mathematics our present exclusive immersion in the formal axiomatic method.'' Using computational methods, previously inaccessible, yet fundamental, questions are now becoming amenable to analysis.
Complementing the above work, I have also pursued experimental approaches to understanding human cooperation and altruism. In some recent experiments we have shown that the altruistic choices of human subjects are economically rationalizable in the sense that they could be accounted for by a standard model of economic choice. We also found that while the altruistic preferences of subjects fall into three distinct classes, there is a large degree of heterogeneity in their preferences.
At the conclusion of Origins of the Species, Darwin remarked that ``[T]here is grandeur in this view of life, with its several powers, having been originally breathed into a few forms or into one...from so simple a beginning endless forms most beautiful and most wonderful have been, and are being, evolved.'' Through a remarkable convergence of ideas, technology, and scientific and engineering imperative, we now stand ready to exploit this ``view of life'' in a quest to understand and control complex adaptive social systems.
There is grandeur in this view of social life.
Some Projects
Ex Machina: Coevolving Machines and the Origins of the Social Universe
If we could rewind the tape of the Earth’s deep history back to the beginning and start
the world anew, would social behavior arise yet again?
While the study of origins is foundational to many scientific fields,
such as physics and biology, it has rarely been pursued in the social sciences.
Yet knowledge of something’s origins often gives us new insights into the present.
In Ex Machina, I introduce
a methodology for exploring systems of adaptive, interacting, choice-making agents,
and uses this approach to identify conditions sufficient for the emergence of social behavior.
Using ideas from biology, computation, game theory, and the social sciences, I explore how
evolving automata can go from asocial to social behavior, and find that systems of simple
adaptive agents—seemingly locked into an asocial morass—can be rapidly
transformed into a bountiful social world driven only by a series of small evolutionary changes.
Such unexpected revolutions by evolution may provide an important clue to the emergence of social life.
Complex Adaptive Systems: An Introduction to Computational Models of Social Life
Scott Page (Michigan) and I have written a book on complex adaptive social systems
(Complex Adaptive Systems, Princeton University
Press, 2007).
In the book, we derive some fundamental examples and principles
of complex adaptive social systems using a simple set of related
models. Models composed of thoughtful agents fundamentally differ from the
typical interacting particle systems that have been widely discussed in
the past. The simple models we develop allow us to
to illuminate, and often correct, some of the key principles that have been
discussed over the past two decades. Finally, we take the opportunity
to outline more fully the foundations necessary for successful
computational modeling in this area.
Discovering Novel Chemotherapies
There is good reason to think that cocktails composed of a variety of chemotherapy drugs might be a key means by which to fight diseases like cancer. Unfortunately, discovering effective cocktails is difficult given the underlying combinatorics. If you have, say, twenty drugs that you could include in the mix, you can create over one million different cocktails. Alas, you can only screen around twenty or so cocktails every few days (laboratory biology is slow and hard), so there is no way to exhaustively search all of the possible combinations. To surmount this problem, we employ a nonlinear search algorithm to direct the discovery of the cocktails. Our preliminary results, focusing on lung carcinomas, are quite promising---the algorithm quickly found a very effective cocktail that is promising enough to begin more elaborate trials. [abstract][pdf]
This work uses laboratory studies of human subjects to show that their altruistic choices are economically rational, in the sense that such choices could result from the rational maximization of a well-behaved (in the economic sense) utility function over giving to self and to others. [abstract][pdf]
Computational Modeling
One of the most powerful tools arising from complex systems research is a set of computational techniques that allow a much wider range of models to be explored. Using these tools, worlds of heterogeneous, adaptive agents can interact in a dynamic environment subjected to various limits of time and space. Having the ability to investigate new theoretical worlds obviously does not imply any kind of scientific necessity or validity---these must be earned by carefully considering the ability of the new models to help us understand and predict the questions that we hold most dear.
Over the last decade or so, computational modeling is beginning to gain a foothold in theoretical social science. Like any theoretical tool, great care and effort is required to create high quality models using computational methods. Here are some links focused on computational modeling in the social sciences (including our graduate workshop). Part of our book on complex adaptive social systems will also be devoted to this topic.
Selected Research Documents
John H. Miller, Ralph Zinner, and Brittany Barrett, ``Directed Discovery of Novel Drug Cocktails,'' Santa Fe Institute Working Paper 05-07-031, (2005). [abstract][pdf]
Jose Lobo, John H. Miller, and Walter Fontana, ``Neutrality in Technological Landscapes,'' Working Paper, 2004. [abstract][pdf]
Ken Kollman, John H. Miller, and Scott E. Page, ``A Simplified Framework for Analyzing the Behavior of Political Institutions,'' 1999.
John H. Miller, ``A Genetic Model of Adaptive Economic Behavior,'' University of Michigan working paper, 1986. This was, to my knowledge, the first paper in economics to use genetic algorithms. [abstract][pdf]
John H. Miller and Scott E. Page, Complex Adaptive Social Systems: An Introduction to Computational Models of Social Life, Princeton University Press, 2007.
Ken Kollman, John H. Miller, and Scott E. Page (eds.), Computational Models in Political Economy, MIT Press, 2003.
Theodore Bergstrom and John H. Miller, Experiments with Economic Principles: Microeconomics 2nd ed, McGraw Hill, 2000 (1997 1st ed).
Theodore Bergstrom and John H. Miller, Instructor's Manual for Experiments with Economic Principles: Microeconomics 2nd ed, McGraw Hill, 2000 (1997 1st ed).
John H. Miller and Scott E. Page, ``The Standing Ovation Problem,'' Complexity, 9 (2004):8-16. [abstract][pdf]
John H. Miller and Scott Moser, ``Communication and Coordination,'' Complexity, 9 (2004):31-40. [abstract][pdf]
John H. Miller, Carter Butts, and David Rode, ``Communication and Cooperation,'' Journal of Economic Behavior and Organization, 47 (2002):179--95. [abstract][pdf]
John H. Miller, ``The Coevolution of Automata in the Repeated Prisoner's Dilemma,'' Journal of Economic Behavior and Organization 29:1 (January, 1996):87--112. [abstract][pdf]
James Andreoni and John H. Miller, ``Auctions with Adaptive Artificial Agents,'' Journal of Games and Economic Behavior 10 (1995):39--64. [abstract][pdf]
John Rust, John H. Miller, and Richard Palmer, ``Characterizing Effective Trading Strategies: Insights from a Computerized Double Auction Tournament,'' Journal of Economic Dynamics and Control 18 (1994):61--96. [abstract][pdf]
John Rust, John H. Miller, and Richard Palmer, ``Behavior of Trading Automata in a Computerized Double Auction Market,'' in D. Friedman and J. Rust (eds), The Double Auction Market: Institutions, Theories, and Evidence, Addison Wesley (1992):155--98. [abstract][pdf]
John H. Holland and John H. Miller, ``Artificial Adaptive Agents in Economic Theory,'' American Economic Review, Papers and Proceedings 81 (May, 1991):365--70. [abstract][pdf]
John H. Miller, ``The Evolution of Automata in the Repeated Prisoner's Dilemma,'' in Two Essays on the Economics of Imperfect Information, Ph.D. Dissertation, University of Michigan, 1988.
Computational Political Economy
Kollman, John H. Miller, and Scott Page, ``Consequences of Nonlinear Preferences in a Federal Political System,'' in Diana Richards (ed.), Political Complexity: Nonlinear Models of Politics, University of Michigan Press, Ann Arbor, MI (2000):23--45.
Ken Kollman, John H. Miller, and Scott E. Page, ``Decentralization and the Search for Policy Solutions,'' Journal of Law, Economics, and Organization, 16 (2000):102--28.
Ken Kollman, John H. Miller, and Scott Page, ``Political Parties and Electoral Landscapes,'' British Journal of Political Science 28 (1998):139--158. [abstract][pdf]
John H. Miller and Peter Stadler, ``The Dynamics of Adaptive Parties under Spatial Voting,'' Journal of Economic Dynamics and Control, 23 (1998):171--189.
Ken Kollman, John H. Miller, and Scott Page, ``Political Institutions and Sorting in a Tiebout Model,'' American Economic Review 87:5 (December, 1997):977--992. [abstract][pdf]
Ken Kollman, John H. Miller, and Scott E. Page, ``Computational Political Economy,'' in Brian Arthur, Steven Durlauf, and David Lane (eds.), The Economy as an Evolving Complex System II, Addison Wesley, 1997.
Ken Kollman, John H. Miller, and Scott Page, ``Adaptive Parties in Spatial Elections,'' American Political Science Review 86 (December, 1992):929--37. [abstract][pdf]
Experiments in Cooperation and Altruism
James Andreoni and John H. Miller, ``Giving According to GARP: An Experimental Study of Rationality and Altruism,'' Econometrica, 70 (2002):737--753. [abstract][pdf]
James Andreoni and John H. Miller, ``Rational Cooperation in the Finitely Repeated Prisoner's Dilemma: Experimental Evidence,'' Economic Journal 103:418 (May, 1993):570--85. [abstract][pdf]
Stephen Lansing and John H. Miller, ``Cooperation, Games, and Ecological Feedback: Some Insights from Bali,'' Current Anthropology, 46 (2005). [abstract][pdf]
John H. Miller, ``Evolving Information Processing Organizations,'' in Alessandro Lomi and Erik R. Larsen (eds.), Dynamics of Organizations: Computational Modeling and Organization Theories, MIT Press, Cambridge, Massachusetts (2001):307--27. [abstract][pdf]
John H. Miller, ``Active Nonlinear Tests (ANTs) of Complex Simulations Models,'' Management Science 44:6 (June, 1998):820--30. [abstract][pdf]