IlliGAL Blogging

I want to comment on Paul Winward's brief post regarding ISGEC becoming ACM's newest SIG, SIGEVO. I've been heavily involved in the process of this transformation, and so my views are not entirely unbiased (no one's views are entirely unbiased or so any good postmodernist would claim), but I think this move is good for ISGEC, good for ACM, good for the field of genetic and evolutionary computation, and good for computer science. For too long, GAs and EC more generally have lived in the backwaters of academic and industrial life. On the one hand, it is hard to understand why this is. After all, evolution is one of the biggest ideas in all of science, and the impact of modern genetics is hard to overstate, so it seems quite natural that computations based on evolution and genetics should have an important place in science and technology. Why, then, have GAmists and ECers been largely relegated to second tier schools, non-tenured posts, or non-CS departments (even, god forbid, engineering departments)?

To some extent it's a matter of bad timing. First wave evolutionary computation researchers (and cyberneticists of all stripes) in the late 50s and 60s were just getting their sea legs under them when Minsky and Papert's premature and misleading hatchet job on neural networks appeared and took the wind out of cybernetic sails generally. Moreover, computers of the day weren't entirely up to the tasks put to them, and a number of prominent pioneering GA/ECs works were prematurely dismissed as no big deal. Finally, the rise of symbolic AI and the subsequent cybernetics winter made it academically disreputable to persist in the "folly" of genetic algorithms and evolutionary computation, but persist the field did, and we are here today because of the courage of a small group who swam against the intellectual currents of the 60s, 70s, and 80s.

Another reason for the current situation is that human lifespans are no longer a good match to the turnover in intellectual ideas. This brings to mind the old story told of physicist Max Plank. When he was asked how the acceptance of his revolutionary ideas in quantum physics was going, it is reported that he said, "Wonderful, lots of funerals." In Plank's day, perhaps the natural turnover in faculty matched the turnover of ideas, but in our time, rapid change in thinking has not been matched with a concomitant shortening of faculty life expectancy. The result is large numbers of powerful faculty in charge with ideas in mind that are more than a little behind the time. Unfortunately, the most obvious solutions to this problem are serious felonies, and no one is here suggesting that steps be taken to cull the herd.

So, it is in this sense, that the acceptance of SIGEVO by ACM couldn't come at a better time, and should be seen as a very positive thing. With GECCO as an ACM conference proceedings, young faculty can safely put their work there without endlessly defending their choice. With ACM affiliated journals in EC and GP, department heads have the moral equivalent of a Good Housekeeping seal of approval on our field to rely upon instead of the pleas of a lone faculty member up for tenure. More importantly, if SIGEVO is seen as part of ACM and CS, it will become easier for CS department heads to hire faculty with GA/EC credentials.

Of course, department heads when asked about tenure and promotion will tell you that each case is special and is scrutinized without regard for brand names. Department heads have told me that the decision to move from ISGEC to SIGEVO will make absolutely no difference to them. But consumers will tell you that their decisions to buy products are also made on a case-by-case basis without regard for brands--as their grocery baskets fill with Crest, Tide, Mach 3, Ragu, Pepsi, Twinkies, and Tide. In a busy world, trusted brands allow buyers to get quality products at low risk of error and low search costs. In the new world of ACM and SIGEVO, CS departments and their heads will be able to rely on a trusted brand in one of the most important decisions they make, the hiring, retention, and promotion of faculty.

It will take some time to know the overall effect, and indeed, joining a large bureaucratic organization like ACM will come with its share of constraints and costs. But, in the end, I believe the young people in the PhD pipeline will benefit immensely through higher probability of hiring, better chances to tenure, and improved prospects of funding as the once outlaw field of genetic algorithms and evolutionary computation comes in off the range, hangs up its six-shooter, and becomes a law-abiding denizen of the mainstream of computer science.