Analogical Reasoning in Machines

Posted 15 Aug 2005 at 16:54 UTC by steve

Identifying pairs of words that are analogous is easy for humans. For example, most of us can easily understand that the words cat:meow are analogous to dog:bark. Our ability to identify the relational similarity between simple pairs of words is thought to underly many cognitive and linguistic processes. A new paper by Peter D. Turney titled Measuring Semantic Similarity by Latent Relational Analysis (PDF format), describes an algorithm called, you guessed it, Latent Relational Analysis (LRA) that can give machines the ability to measure the relational similarity of words. The algorithm uses a vector space model (VSM) in which words are represented by vectors determined by the frequencies of patterns. The similarity of words is calculated from the cosine of the angle between two word's vectors. The new algorithm was able to achieve human-level performance on analogy questions from a college-level multiple-choice test. For more information see the author's website.

Cosine of 42, posted 17 Aug 2005 at 12:21 UTC by The Swirling Brain » (Master)

Cosines are the answer to everything AI? It's a novel approach but I really don't believe cosines of vectors applied to words is really the way to associative learning. But who really knows all the strange ways the brain computes things? A mind would have to really "understand" the association for the program to really work or else some associations would work and some wouldn't and therefore be an intermittent method. I still feel that there are a whole lot of writing about AI and not too much doing in AI. If only everyone were a gung ho computer scientist that didn't care too much about design (ood) and just loved to compose quicky programs on a computer. Perhaps then there would be enough computer monkeys to tap out some real AI code. Instead we are left with philosophical paper writers that just hem-haw and beat around the bush about how they are the ones who really know what AI yet can't produce. I guess you do have to talk about it and thrash it out before you can make one. But weird illogical stuff like this really does make one long for the future day when we really see a smart robot. At that time, we'll look back at all the wacky ideas and wish we lived back in simpler times?

Pining for days when computers had toggle switches, posted 17 Aug 2005 at 12:49 UTC by jeffkoenig » (Master)

>At that time, we'll look back at all the wacky ideas and wish we lived back in simpler times?

Prolly. I'm now kicking myself for throwing out all my issues of "Creative Computing" and "Byte". The ads for both software and hardware would be a hoot.

See more of the latest robot news!