I found this interesting site by doing a no-no... I clicked on one of the ads that Google placed on my own site... I wasn't supposed to, but it looked interesting to me, so I broke the rule. At this point I am glad I did.
Monica Anderson has created a terrific little site where she presents her ideas concerning a somewhat unconventional viewpoint on and approach to machine learning.
She posits very articulately and with powerful logic, that logic processing is not the most important underlying property of intelligence. That in fact many things are quite intelligent without being the least bit logical. Intuition, she asserts, is more foundational, more important, and quite possibly will be more fruitful as an approach to AI development efforts.
She summarizes it quite well:
"Most humans have not been taught logical thinking, but most humans are still intelligent. Contrary to the majority view, it is implausible that the brain should be based on Logic; I believe intelligence emerges from millions of nested micro-intuitions, and that Artificial Intelligence requires Artificial Intuition."
Designed to be read more or less straight through, the site is essentially a long essay or short treatise on her ideas. Though the site design camouflages this fact smoothly by breaking it into sensible page divisions and is really quite excellent overall.
I personally am particularly intrigued by anyone that is talking about a new approach to or viewpoint on AI development. Since it appears to me that there is no other explanation for the "AI Winter" we have been in for 2 decades except that we have over-committed to some foolish and fruitless research pathways that need to be abandoned in favor of some fresh ones. (OK, I know there really has been no "AI Winter" and that machine intelligences are in fact growing rapidly in number, power and ubiquity, but since I can't yet go on the web and talk to HAL, I am not happy with the current state of AI development so I reserve the right to speak of an AI Winter until I can kick it online with HAL.)
This is probably a little more of her site than I should block quote, but she makes such a very powerful statement here and it resonates so much with my thinking on AI - specifically on the reasons AI is not bearing more interesting fruit - that I just had to include it:
"The hard sciences, such as Mathematics and Physics, insist on correctness. Computer Science was born in Mathematics departments at universities worldwide, and Computer Science is therefore a hard science. Programs are expected to be correct and to run as specified. Artificial Intelligence was born in Computer Science departments, and inherited their value sets including Correctness. This mindset, this necessity to be logical, provable, and correct has been a fatal roadblock for Artificial Intelligence since its inception.
The world is Bizarre, and Logic can not describe it. Artificial Intuition will easily outperform Logic based Artificial Intelligence for almost any problem in a Bizarre problem domain.
From the very beginning, Artificial Intelligence should have been a soft science."
She discusses very convincingly a framework for considering what she calls Bizarre Domains. She outlines the problem space effectively and populates it with a number of things I had not considered collectively before.
She talks about holistic approaches, paradoxes and strange loops - all of these things give fits to conventional Artificial Logic development.
As I was reading along I began to think, she should really read 'Blink' by Malcolm Gladwell. Her advocacy of intuition over logic reminded me strongly of his ideas... In the next paragraph she refers to the book. She also recommends Godel, Escher Back and Zen and the Art of Motorcycle Maintenance as suggested reading - two of my all time favorites. So it seems I have found something of an intellectual kindred spirit.
Her research and evangelism seem to be focused mainly upon using artifical intuition to extract semantics from language. (Which is a primary problem in AI development, one that has not borne the fruit I want to see) She maintains algorhythmic details as a trade secret and will not discuss them except under NDA, so I cannot comment on the math, but reading this site has convinced me that the notion of Artificial iNtuition (AN) is more basic than artificial logic and that it is in fact a necessary foundation upon which any operable strong AI will need to be built.
I will continue to read up on these ideas. I recommend this site highly (It can be read completely in an hour)
Thank you for the lovely comments about my site. Kindred spirits indeed; we are using insights from "I am a Strange Loop" (which you mention elsewhere) to guide our research. And it sounds like you have read Tor Nørretranders: "The User Illusion".
ReplyDeleteThe AN site is over two years old. I have since figured out a few more things and gotten better at telling the story. I have also gotten bolder. My focus groups told me to avoid the H-word on the AN site. Nowadays I'm Full Frontal Holistic. And this is not your aunt's Crystals and Aromatherapy Holism, this is Epistemological Double Barrel "Context is Everything" Holism.
I have given many public talks on these issues. If you want to hear my latest insights, go to http://videos.syntience.com and watch the videos there. Hours of fun if your mind is sufficiently open. Don't miss the talk by Peter Norvig, Director of Research at Google, since what he says has much bearing on what I've been saying; and we use the word segmentation test he talks about daily in our research. Syntience Inc. and Google are not the only companies pushing in this direction, but I think we're leading the charge.
If you are ever in Silicon Valley and are willing to sign an NDA, please get in touch and we'll tell you how the implementation works. This offer is available to anyone that can provide a plausible argument that there can be mutual benefit to the exchange (such as "funding", but not limited to that).