 |
|
|
|
|
|
||
|
|
||
|
|
||
|
|
|
|
|
|
||
|
|
||
|
|
||
|
|
|
|
|
|||||||||||||||||
Add Forum to Favorites | Send Topic To a Friend | View Forum FAQ | Track this topic |
Last Thread Next Thread |
| The Intuitive Algorithm |
|
 |
Anonymous Poster |
|||
 Posted - 7/1/2001 11:54:29 AM |
||||
| Has anybody else read this article? If so, I would like to discuss it. |
||||
|
||||
|
Anonymous Poster |
|||
 Posted - 3/10/2002 1:36:22 AM |
||||
| Does anybody know how to contact the author? i'd like to ask a few things to him, and maybe after about 4 years later ha may have new ideas on this. |
||||
|
||||
|
Anonymous Poster |
|||
Posted - 11/17/2004 12:47:26 AM |
||||
| The "Intuitive Algorithm" essay was written by me. My e-mail address is abraham1@vsnl.com. A book with the same name has been published and the publisher's e-mail address is ewb@touchtelIndia.net |
||||
|
||||
|
Anonymous Poster |
|||
Posted - 3/23/2005 4:15:28 PM |
||||
| Very, VERY, good essay! |
||||
|
||||
|
Morbo Member since: 5/15/2000 From: Canada |
|||
|
Posted - 3/23/2005 5:44:33 PM |
||||
| While I haven't finished the entire article yet, I already have my doubts about it. So far, in over 4 scrolled pages, all I've read about is about how some undefined algorithm is going to solve a massively parallel problem on a serial processor 'instantly' seamingly just by defining the problem as a spreadsheet. I don't have time to finish reading it yet, but a cursory scan over the rest of the article fails to locate any details on how this is to be accomplished. The algorithm appears to be nothing more than a tri-state expert system. Though the article claims it's 'far more advanced than any other expert system', the only description it's offered up thus far is nothing that I haven't worked with before (in the process control industry, which has some very powerful expert systems that are implemented on custom hardware). Without getting into a critique of the quality of the writing, so far all I've read is a series of problems and how this nebulous algorithm solves it 'instantly', with no further explanation or detail. The lack of a specific, concise introduction, as well as continuing statements about how great this algorithm is without offering any proof or details, doesn't lend to it's crediability one bit. While it does appear to compile a great deal of modern information on brain physiology and psychology, I'm not very confident it will lead anywhere. For now, it appears to simply spell out a lot of solutions that have already been invented, without knowing they have been. I'll check back once I managed to finish it all, hopefully it manages to prove me wrong. |
||||
|
||||
|
Anonymous Poster |
|||
Posted - 5/26/2006 12:51:12 AM |
||||
| I have problems with the concept that I have made exaggerated claims for the Algorithm. The idea is simple. Every disease in a group of diseases is coded Yes/No/Uncertain for every symptom question. "Yes" answers elminate diseases which lack the symptom. No elminates diseases which have the symptom. Uncertain take no action. So, instead of serial searches, with the problem of exponential growth, the algorithm elminates diseases for every question and quickly arrives at answers. I suggest neurons operate similarly. They recognize inputs and switch off if they do not recognize a pattern of inputs. Inhibition is well documented. When large numbers of questions can be asked in parallel, as in the parameters of a power plant, speedy answers are possible. For the nervous system, millions of simultaneous inputs occur, creating instant recognition. A face is recognized, because a single feature is unique among all known faces. So all other possible faces are eliminated. Your recognize a friend, with virtually no time lapse. Is this rocket science? My website is www.intuition.co.in. You can test the logic of a sample expert system there too. Abraham Thomas |
||||
|
||||
|
Anonymous Poster |
|||
Posted - 10/15/2006 12:43:27 AM |
||||
| I think the author would do well to have a look at his example with the white and black swan and reconsider the merit of using logical exclusion in the human mind. I'll just quote the end; "In a similar manner, IF A SYMPTOM IS ALWAYS PRESENT for a particular disease, inductive logic also implies that an absence of the symptom excludes that disease from further consideration." I boldfaced the extremly important assumption that is REQUIRED to make exclusion a logically valid technique. The human mind unfortunatly deals with a wide variety of problems and can rarely make this assumption and expect to reach universally useful conclusions. Especially because exclusion is hard to combine with creativity and expansion of established concepts. Consider a person that has only seen black and white soccer balls. It would be natural that his internal definition of soccer balls would include all the physical characteristics of the ball. Including color. Now, what do you do when you see your first blue and green soccer ball? Do you bluescreen? No, you go "hey, that's a soccer ball, only it has different colors! Cool!". What if you see an elephant that has had its trunk cut off? "ZOMG! A new species!"? I don't think so. This indicates that the image->concept algorithms are, at the very least, capable of an additional type of search that returns results that don't match completely And can do this almost as fast as the regular search, but a more likely explaination is that exclusion is not used as a general search strategy when mapping concepts. It could potentially be used elsewhere though. It's also pretty interesting that the author thinks that concepts free will and viewing the brain as a collection of finite state machines are compatible. Personally I'm more in the loop-motivated automaton camp. Other than that the article was an interesting read. |
||||
|
||||
|
Abraham Thomas Member since: 10/24/2004 From: Chennai, India |
|||
|
Posted - 6/15/2007 1:55:57 AM |
||||
| Happy that the writer found the article interesting. But, the soccer balls example ignores the scale of the stores of wisdom in the mind. As an example, the data stores in DNA in the body are reported to be sufficient to fill the Grand Canyon 50 times over with 500 page code books. Recognition also uses a galactic data base of combinatorial memories in one hundred billion nerve cells. Cell memories evaluate tastes, smells, sounds, visual images and millions of remembered situations within a few milliseconds to trigger recognition of the ball. Size, shape, stadium sounds, context are all perceived patterns. Only a system, which swiftly eliminates the impossible, can act so fast. Cells, which recognize golf balls, tennis balls, or ships become inhibited. With the wisdom of the mind, a green soccer ball in a game becomes new information. If “free will” implies a conscious choice, it does not exist. Libet's experiments showed that motor systems began to function 350 milliseconds before a subject believed he “willed” the pressing of a button. There is no way you could evaluate a million possibilities before you choose between eating and drinking. Your system makes those choices. Abraham Thomas |
||||
|
||||
|
radagaisus Member since: 6/1/2008 |
|||
|
Posted - 7/11/2008 2:43:41 PM |
||||
| You can explain this easily if the conscious part is only a part of your brain, and after you receive an image of a bull running your way your body will start pumping adrenaline. Yet I don't know what do you mean by 'motor systems'. Also, it is a smart strategy evolutionary for the brain to have several 'think' algorithms. You don't control your reflexes because maybe they are a different part of your brain, stuff that have been proven so good or so important that you don't have to think before you do them. Also, about that soccer ball thing. Your pattern recognition algorithm should identify if something is elementary for a pattern or if it is secondary and might imply a pattern. A soccer ball is still a soccer ball because color isn't his primary attribute but.. size, the feel he has from air pressure inside. This are the important things. But how do you identify what is essential to being 'soccer ball' and what isn't if you are developing AI, I don't know yet. |
||||
|
||||
|
smitty1276 GDNet+ Member since: 1/17/2001 From: Seattle, WA, United States |
|||
|
Posted - 7/11/2008 3:02:31 PM |
||||
Wow... you guys realize take time to chew on and process each others comments, huh?  |
||||
|
||||
|
Abraham Thomas Member since: 10/24/2004 From: Chennai, India |
|||
|
Posted - 9/1/2008 2:20:21 AM |
||||
| Radagaisus suggests that there could be several think algorithms. It is a single algorithm, which makes holistic communication possible in the network. When an animal chooses to chew grass, or drink water, multiple intelligences, including motor neurons, recognize, feel, select options and deliver motor output; all, within milliseconds. Independent think algorithms would create both conflicts and communication problems. To reach a final decision, the intuitive algorithm uses elimination by every neuron to handle all decisions, including uncertain ones - two faces, or a vase. The algorithm enables each one of a hundred billion neurons in the interactive system to make a single final choice from a billion possibilities. |
||||
|
||||
|
Abraham Thomas Member since: 10/24/2004 From: Chennai, India |
|||
|
Posted - 9/17/2009 12:36:03 AM |
||||
| Prof. Marvin Minsky, of AI Labs at MIT commented on my book "The Intuitive Algorithm." "Thank you for the book! ..... I like both your ideas and your style; every small chapter has some valuable concepts, along with well-designed examples! As it happens, I generally do not agree with the classic distinctions of feeling from thinking -- but here that may be just a matter of words, because I think that you have well described many aspects of how our minds manage to deal with the world. Similarly in many ways, your theories are somewhat different from mine - but many of those differences come from my experiences with programming. In fact I wholly agree with you that the logical Prolog-like approach does not reflect much mental reality. Indeed, I am convinced that Analogy is what our minds do best....." Maybe the ideas make some sense! Abraham Thomas :) |
||||
|
||||
|
Compugasm Member since: 10/18/2009 |
|||
|
Posted - 10/18/2009 7:31:25 PM |
||||
| A lot of people are confused by the example of medical diagnosis from the original author. The simple analogy, for us internet folk, to the exclusion argument is Google Search. When you start typing in the search box, any patterns (a pattern being the letter and words as you type them) are eliminated from the set of possible outcomes. In effect, every character you type into the search box is a Y/N/U question being asked of the expert system. The matching patterns are displayed in a drop down box below what you are typing. As soon as enough patterns have been eliminated, you can simply select a remaining pattern, without having to ask further questions. That's it. Since we all are aware of how fast and accurate Google is, there should be little doubt in the authors claims regarding the speed and accuracy of medical diagnosis when using the Intuitive Algorithm. I mean, all you have to do to test the claims, is to use Google. [Edited by - Compugasm on October 18, 2009 7:31:02 PM] |
||||
|
||||
|
swiftcoder Member since: 7/3/2003 From: Boston, MA, United States |
|||
|
Posted - 10/18/2009 8:10:46 PM |
||||
<offtopic>This particular discussion thread suffers from the worst case of thread zombie-ism I have ever seen. Seriously, will it never die? The article is from '97, folks! And we won't even mention that half the posters in this thread have only ever posted in this thread...</offtopic>  Tristam MacDonald - swiftcoding |
||||
|
||||
| All times are ET (US) |
Last Thread Next Thread |
|
