Christoph Adami, Richard K. Belew,
Hiroaki Kitano, and Charles E. Taylor, Editors
MIT Press, 1998
Back to Table of Contents
It was fifty years ago in 1948, on the occasion of the Hixon Symposium at Caltech, that John von Neumann gave his celebrated lecture entitled "The General and Logical Theory of Automata" (von Neumann vn), where he introduced to the public his thoughts on universal, self-reproducing machines. Von Neumann himself professed to have been inspired by Turing's theory of universal automata, which dates back another ten years. Also at the symposium was Warren McCullough, who five years earlier had introduced, with Pitts, the universal neuron which is at the center of almost all work on artificial neural networks. This work was, of course, also founded on Turing's ideas about universality.
For fifty years then, have researchers fought to capture the universality in Life, to transplant it into a different medium, and study its "general and logical" characteristics. These intervening years have been sometimes quiet, sometimes punctuated by feverish activity. Clearly, Chris Langton's work and the Artificial Life conferences that he spawned in 1987, have ushered in a new epoch of ALife research. Many consider Tom Ray's "tierra" system (introduced at the second ALife Workshop [Ray1991]) another major milestone, arguably synthesizing a truly living system within a computer.
These are the Proceedings of the Sixth International Conference on Artificial Life. Like all things alive, the conference too has mutated and adapted over the years, from a ``Workshop on the Synthesis and Simulation of Living Systems'', attended by 150 researchers in a wide variety of fields, to the International Conference attended by many times more.
The theme of this year's conference, "Life and Computation: the Boundaries are Changing" addresses two of the topics touched upon above. First, von Neumann and Turing taught us that Artificial Life has its root in the universality of computation extended to the universality of life. But even these visionaries would be amazed at the mind-boggling variations on computer/life interactions represented in here: molecules used for computing, computers modeling molecules, self-assembly in thin films, resurrected fossils, evolving programs, statistical models of genetic populations, robotic crickets, developmental and immunological models, social and linguistic models, artificial architecture, the economics of agents. What would they have thought?!
Indeed, the field does not stand still: it is itself evolving, and the boundaries are changing. For example, there are more papers dealing with computational molecular and cellular biology at this Alife meeting than ever before, providing new insights into developmental processes in the fruit fly, mechanism for cell-differentiation, as well as the modeling of immune-reponse. And while the emergent properties of agents have always been a mainstay of the Alife field, today we are witnessing increasing applications to financial markets, trading, and even Internet transactions.
But we also are witnessing efforts at reshaping some of the staples of what used to be Artificial Life research. Stuart Kauffman's NK-model [Kauffman and Levin1987], has been used to model the ``ruggedness or smoothness'' of evolutionary landscapes for over a decade. But Barnett introduces in these proceedings an ``NKp''-model with a form of neutrality that many believe (e.g., see [Eigen1986], for a forceful argument) essential to evolution. Similarly, Chris Langton's model of self-replication as virtual state machines on a Cellular Automaton (Langton langton84, langton86) is a second classic reference in Alife. Sayama introduces in these pages a form of `death' into Langton's model, transforming the crystalline but abiotic structrues into much more life-like forms that even seem to evolve. Both of the latter presentations are outgrowths of Master theses: a sign that Alife is continually renewing.
As always, there is much more worthy work going on than we can publish in any one Proceedings volume. We received about 100 submissions and 39 of these are presented as full papers. We have highlighted nine of the papers as exemplary of high-quality work across the entire spectrum of ALife topics; we can recommend these especially to readers trying to see just what ALife is about in 1998. In addition, we include 22 shorter papers (which will be presented as posters during the meeting). Many of these extend ALife into exciting new directions, or bring a new student's or scientist's perspective.
We are especially excited by the set of invited speakers that have agreed to participate in ALife VI. Christos Papadimitriou (UC Berkeley) and Len Adleman (USC) are seminal computer scientists, wrestling with many of the same issues that concerned Turing and von Neumann but benefiting from the great progress in computational complexity analysis developed in the interim. Gerald Joyce (Scripps Research Inst.) has been an active participant in ALife work for a number of years and combines an understanding of the computational issues with a practioner's insight into what is possible in a test-tube (while succeeding even at ethings that seem impossible!)
Beyond the keynote speeches, the plenary talks, the parallel sessions and the posters which are covered in these proceedings, Alife VI sported events that remain undocumented, such as the Alife Art Show, workshops, demonstrations, a robot contest, all occurring right at Hollywood's doorstep. In the shadow of a city that many brand `artificial', the `Alife experience' remains real and we have come full circle. Fifty years after Caltech hosted the Hixon Symposium, Southern California again provides a nurturing environment for Artificial Life. On to the next fifty years!