Go to: Home page of www.BillHowell.ca
Holidays - Neural Networks and Genomics
Celebrating 20 years of neural networks!
IJCNN2007 marks the 20 year anniversary of the International Joint Conferences on Neural Networks. Please do contribute to and participate in this event and celebrate this important milestone of the neural networks community with us!
Thank you, and we look forward to seeing you in Orlando!
Arizona State University
Rensselaer Polytechnic Institute
IJCNN04, Budapest Hungary (was www.conferences.hu/budapest2004) - Last summer's conference in Budapest included a special symposium on retinal prosthesis (artificial retinas for macula degeneration and retinal pigmentosis or however you say these conditions) – which are already in clinical trial in California (for 2.5 years now).
IJCNN05 Montreal (http://faculty.uwb.edu/ijcnn05/) - In Montreal we hope to have special sessions on genomics and the brain (likely from the computational/ modelling perspective), and brain implants (Canada's National Research Council (NRC) is one of many groups looking into these). Initial applications for brain implants would be diseases (like Parkinson's, Alzeihmer's, Lou-Gherig's) and brain damage. Many other major themes will also be addressed (advanced control theory etc – just glance at the special sessions and list of topics on the web page). Fun like you can't believe!
I'm a "Program Co-Chair" for this conference, meaning my nights and weekends that were freed up from less overtime for work are now jammed full of work to help organize the IJCNN05 conference. In other words, it's summer vacation all year long! I find that it's definitely far more scientifically interesting that anything we do in our research labs at work, or I wouldn't be doing this for my vacation!...
BioNorth 2004, Ottawa (www.olsc.ca) - I won't get into details, especially as my focus, and that of the conference this year, was partnering, patents, bio-business, and regulatory, rather than technology. Suffice it to say that I was a little upset to see how slow and problematic the regulatory process is (especially from a business/ investment point of view). No doubt there is some need for caution especially with new and unknown technologies, but it looks like the fictitious "zero risk" (again the omnipresent zero-dimensional thinking) is a reality, rater than a compromise/ trade-off of risks adjusted for the individual.
Running for Governor – International Neural Network Society (INNS- www.inns.org) - Believe it or not, I'm in the running as a candidate for governor of my summer vacation again this year! My objective is to double my support this year – if I can find someone else to vote for me... I'm thinking of hiring a Liberal party vote-buyer to bolster my campaign <grin>.
(Dec03 Xmas letter)
I have two types of vacation – most importantly to visit my family all over Western Canada, then to attend the International Joint Conference on Neural Networks (IJCNN) annual conferences. I also try am very interested in genetics, but can only afford to attend the local annual conference (given a host of other spending priorities).
IJCNN02, Waikiki Hawaii, July02 - The Waikiki conference was a real blast – really three conferences in one (neural networks, fuzzy systems and evolutionary computation). I spent 8 days in Hawaii, and even though I love swimming in the ocean and running, I only managed to go jogging once, do quick swims twice, plus take a half day bus tour/ swim at ?Huanamo? bay (fish, turtles galore! - I rented flippers, mask and snorkle and had a beautiful swim! plus scarred my belly flopping around on top of coral like a beached whale). The rest of the time I spent in the conference hotel, where the real fun was... Three "fun highlights" of the conference were:
Harold Szu's presentation on information theoretic temperature and the derivation of Hebbian learning from first principles - That may or may not mean much to you, but Szu is a real fun guy to be around, the theory is a real thought provoker, and Hebb did much of his work at McGill University's Neurological Institute in Montreal say 40 or 60 years ago. Say no more...
David Fogel – who was the general conference chair, twisted my arm to buy his new book "Blondie 24" (he and Don Wunsch also got me to volunteer to do visa letters, as mentioned below). Blondie is an evolutionary neural network, which, when given only minimax theory and the rules of checkers, becomes an expert level player (but not a grandmaster!) in two or three months. It's an extraordinary accomplishment, given that no "coaching, expert rules, or hints/ heuristics" were given to Blondie (as was the reason for Big Blue's defeat of Kasparov in chess – which was the inspiration for Fogel to create Blondie). Fogel sees Big Blue as being a sideline on the road to intelligent systems, and his belief is that intelligent systems MUST be evolving systems (learning, training, parameter adjustments are not sufficient).
Robert Hecht-Nielson – was really excited about the breakthrough potential for his computational model of the thalamo-cortex. I like the way that he put it (which irritated everybody else who thinks that THEY know how the brain works):
"...Our advice is to start your research immediately, run as fast as you possibly can, and never slow down. In a very short time you will hear the starter's pistol fire in the distance behind you, releasing the largest intellectual land rush of all time...[Hecht-Nielson]"
IJCNN03, Portland Oregon, endJuly03 - For the Portland conference, I spent an enormous amount of time from Oct02 through Jul03 arranging for letters of invitation for international scientists who planned to attend this conference, as tightening security in the USA following 11Sep01 has made it much more difficult to obtain visas even for short stays. There were between 200 and three hundred scientists who needed letters, of whom we know 24 couldn't get the visa in time, and perhaps as many as 100 did not come because of visa challenges (including those that we did not hear from). Anyways, it was a real pleasure to work with the conference organizing committee – Chair Don Wunsch (U Missouri-Rolla), Technical Chair Michael Hasselmo (BostonU), Technical Co-Chairs Kumar Venayagamorthy (U Missouri-Rolla) and DeLiang Wang (OhioStateU). But most of all Tulay Yildirim (YildizTechU Turkey) and Heng Hiot Lim (NanyangTechU Singapore), who worked with me to provide letters to help with scientists' visas.
IJCNN04, Budapest Hungary, IJCNN05 Montreal - Next summer's conference in Budapest will include sessions on cochlear and retinal implants (artificial ears and eyes) – which are already available or in the prototype clinical trial stage. In Montreal we hope to have special sessions on genomics and the brain (likely from the computational/ modelling perspective), and perhaps brain implants (Canada's National Research Council (NRC) is one of many groups looking into these). Initial applications for brain implants would be diseases (like Parkinson's, Alzeihmer's, Lou-Gherig's) and brain damage. Weird things are happening down at the nanometer level, so we should see very interesting developments over the next 100 years. Additionally, trends are moving towards my own pet project, which I'll never have time to pursue, Mindcode:
"Given that computer code is used to program computers, then Mindcode is use to ...[Howell]" - the idea is to look carefully at data, procedures, operating systems, and unknown higher-level abstractions coded in the "junk" DNA). The implications are that much more of what you think may be "assisted" by pre-existing structures than you currently think is the case...
Another implication of genetics (not limited to the above concept) is that the "nature vs nurture" debate is probably pure junk, and even perfectly exact DNA cannot reproduce the same brain – the number of synapses vastly outnumber DNA base pairs (1015/ 3*109 ~ 106). Statistical variations in early brain growth (ontogeny) (before learning really kicks in, and certainly for part of the early learning period) mean that, even at a very over-simplistic conceptual level, three processes must be considered! Past identical-twin studies vastly over-estimate the role of the environment, but not of genetics. But it would be extremely difficult to quantify differences due to ontogeny – as this would require detailed three-dimensional determination of neural structures/ synapses.
So remember, as a conclusion to this section, the next time that you are planning a vacation:
"...The purest form of art, beauty and sheer fun, is mathematics...[Howell]"
BioNorth 2002/03&04, Ottawa Ontario - stem cells, genomics, proteomics etc. What a blast! I would like to go to the big USA conferences, but I already spend too much money on neural networks..
My favourite saying from this conference, though, was from three or four years ago, when a scientist from San Diego (who used to be at McMasterU in Hamilton) gave a presentation on aging related diseases. Progress in understanding aging, he said, has been extremely rapid in the last five years, surprising the experts in the field. Anyways, he concluded his talk by saying:
"...Sometime in the not-so-distant future, someone may be giving you a presentation on the same topic [aging]. It is quite possible that he could finish it by saying "You know, we may be the last generation to die." [approximate quote of - I forget his name]"
Boy, was I ever excited, and could only hope that I would make the cutoff point! Imagine my surprise upon learning that not everyone thinks that eternal (or even 500 year) life is necessarily desirable. I explored these reactions in more detail in an unscientific survey, and got some beautiful explanations and viewpoints.
Running for Governor – INNS - Believe it or not, I was drafted to run as a governor for my summer holidays (the State of California was already accounted for, so I couldn't run for governor there – and that would be too much work anyways). I questioned their judgement, but went along knowing there's no sane way that I'll actually be elected. If ever that happens, I'll have to find out what governors do, then rent the film "Being there" (or something like that), which I've never seen. Anyways, at least I'd have a say in how my summer holidays are organized...