The Society for the Study of Artificial Intelligence and Simulation of Behaviour
According to its home page, the Society was founded in 1964, and is the largest Artificial Intelligence Society in the UK. Activities include giving travel grants, organising conferences, public lectures and other events, and publishing the Interdisciplinary Journal of Artificial Intelligence and the Simulation of Behaviour.
You can find out more about the AISB via the home page. My interest here is to point you at its online publications; and you will find these at the end of my links, below.
https://aisb.org.uk — AISB home page.
https://aisb.org.uk/travel-awards/ — AISB travel awards. Only of interest to those working in the UK, I assume; but given the ever-present press for funding, perhaps a lot of interest to them.
https://aisb.org.uk/publications/submission-instructions-for-the-q/ — AISB page on how to write book reviews and conference reports. Intended for those writing for AISB Quarterly, but may be useful to others. In particular, the section on conference reports gives advice on how to structure the report, and would help those set this task for the first time.
www.aisb.org.uk/publications/proceedings.shtml — AISB Convention Proceedings. With online versions of proceedings, symposia and workshops from 2000 to 2006, there’s a wealth of material here, although it isn’t indexed by author. Amongst other goodies are the three volumes of the 2006 Convention on Adaptation in Artificial and Biological Systems; the Third International Symposium on Imitation in Animals and Artifacts, the Symposium on Next Generation Approaches to Machine Consciousness, and 8 other links; through to 2000, which gives us AI in Bioinformatics, AI and Legal Reasoning, Designing a Functioning Mind, and four other links.
www.aisb.org.uk/aisbq/index.shtml. — Archive of the AISB Quarterly. This is a ten-plus-a-few page newsletter of short articles, book reviews, and conference reports. Online copies run from Autumn 2001 to Autumn 2005. Picking a selection of articles at random, we have: Tim Blackwell on generating music by self-organising swarm interactions; John Koza on the “high-return” AI implemented by genetic programming; Joanna Bryson on the role of emotions in modular intelligent control; and W. I. Sellers and G. S. Paul on estimating how fast giant tyrannosaurs could run using models generated by evolutionary robotics. Not to mention the exploits of Father Hacker.
www.aisb.org.uk/aibites/ — AI Bites. One-page explanations of: constraint solving; evolutionary computation; machine learning; theorem proving; computational creativity; inspirations; and search. Useful for someone with some computing experience but no AI. Written by Simon Colton.
http://www.inf.ed.ac.uk/publications/author/simonco.html — Colton’s Web site was down when I tried it, but this link points to reports he wrote while at the School of Informatics in Edinburgh. These include Evaluating Machine Creativity, and Cross Domain Mathematical Concept Formation.
Why Does Biological Evolution Produce Modular Designs?
I found an interesting blog entry by John Hawks, Why are organisms modular?. In it, he writes about a paper by Nadav Kashtan and Uri Alon of the Weizmann Institute, Spontaneous evolution of modularity and network motifs.
Consider experiments on evolving neural nets by genetic algorithm. These typically result in nets that cannot be divided into functional subcomponents. But biological evolution does seem to produce designs with such components. DNA contains genes, but also other kinds of functional element such as the promoters, enhancers, and insulators that control when genes are activated. Proteins also can be divided into different functional regions; and at a higher level, cells are built from organelles, and bodies from organs.
What properties of biological evolution lead it to create modular systems? That’s the question explored by Kashtan and Alon. I’m reading their work at the moment, but in the meantime, I’ll point you at John Hawks’s write-up.
http://johnhawks.net/weblog/reviews/brain/culture/modular_evolution_circuits_kashtan_2005.html — Why are organisms modular?, by John Hawks, Assistant Professor of Anthropology, University of Wisconsin-Madison. Blog entry for 21st September 2005.
www.weizmann.ac.il/mcb/UriAlon/Papers/SpontaneousEvolutionOfModularityAndNetworkMotifs.pdf — Spontaneous evolution of modularity and network motifs, by Nadav Kashtan and Uri Alon, Departments of Molecular Cell Biology and Physics of Complex Systems, Weizmann Institute. Published in Proceedings of the National Academy of Sciences of the USA, Volume 102, Pages 13773–13778.
Teaching Evolution with Sodarace
Back in November 2004, I mentioned Sodarace. Using software developed by this joint venture between the Soda arts company and Queen Mary college, users can build virtual creatures from simulated springs, muscles and other body parts and then race them against competitors. Creatures can also be designed by AI programs — the Sodarace home page shows some examples — and saved in XML for input to the Sodarace world. XML-minded readers might be interested to see what elements a Sodarace XML file contains.
Looking at Sodarace again to see what’s new, I found a site devoted to “sodaA+”. This is Sodarace software for teaching evolution to school students, developed by a team of computer science students from Queen Mary college as part of the Computer Science For Fun project, “raising awareness that computers can be fun, as well as educational”. It’s aimed at GCSE students in UK secondary schools, but would surely be useful elsewhere.
The site includes a lesson plans page. You can download lesson plans for inheritance of genetic components, mutation, natural selection, and genetic engineering. These come with what I assume to be references to the sections in the UK National Curriculum to which they relate. To give an example, the blurb for the natural selection lesson plan says:
Creatures change over time. This lesson plan is directed towards the idea of Natural Selection, whereby the alterations best suited to the environment give rise to greater reproductive success. After a long enough period, creatures, in different environments may have diverged to such an extent that new species have been created. Most changes are not beneficial (but some are) and too many at one time will result in a lack of success. Therefore, the process of evolution is that of the gradual accumulation of small but advantageous changes over a vast time-scale.
sodarace.net/index.jsp — Sodarace.
sodaplay.com/constructor/beta/sodaconstructor.dtd — For XML experts, this is the DTD for Sodarace creatures.
http://www.cs4fn.org/alife/sodaindex.html — Queen Mary Computer Science For Fun Sodarace page.
www.dcs.qmul.ac.uk/sodarace/teaching/projects/sodaevolution/index.html — Home page of sodaA+, the site for teaching evolution with Sodarace.
www.dcs.qmul.ac.uk/sodarace/teaching/projects/sodaevolution/lessonplans.html — The lesson plans.
www.royalsoc.ac.uk/exhibit.asp?id=3529 — Sodarace: humans vs machine intelligence. Popular-science write-up of the Sodarace exhibit in the Royal Society’s Summer 2005 Science Exhibition.
Backward in Kansas
Maybe it’s better to view Human Evolution as a peaked graph, as opposed to an upwardly linear one. We have, in two thousand years, perfected a technology of unimaginable potential, yet mostly we just use it to make animals move their lips in TV adverts. My own view of human development is best demonstrated as follows:
Find one of those ‘Ages of Man’ charts — the one that progresses from a gangly chimp to a fully developed Homo Sapien. Draw a little television remote control in the hand of each figure. Then hold the chart up to a mirror. Surprise! You’re in Kansas.
Quoted from Evolution in Reverse, comedian and writer Rich Hall’s commentary on the Kansas State Senate’s bill to make teaching evolution in public schools a crime. Published in his book Things Snowball.