Comments on: Jeopardy!, IBM, and Wolfram|Alpha

By: myclob

myclob — Fri, 11 Feb 2011 13:56:57 +0000

In order for artificial intelligence to take the next big step forward, it will have to do more than retrieve answers to questions, but make decisions like: what should I do with my time. In order to do this we should build databases of beliefs or conclusions, and then brainstorms reasons to agree and disagree. At the very top level reasons to agree will be added as validation of the belief, and reasons to disagree will be subtracted. Of course reasons will have to be weighted. I believe the most elegant way to do this is to have reasons to agree and disagree with each reason. So that a reason to agree with a reason to agree will give a total of 1/2 of its score to the highest level reason, and a reason to agree with a reason to agree with a reason to agree will only contribute 1/3 of its points to the highest conclusion. I believe with very little work, this system could come to decisions that are more WISE than the average real person. In this way artificial intelligence will become computer wisdom.

By: marraco

marraco — Fri, 11 Feb 2011 12:02:03 +0000

But woframalpha is supposed to be a search engine to get factual data.

By: kmac086

kmac086 — Thu, 10 Feb 2011 21:57:34 +0000

I reccomend gapminder.org for countries income, gdp, ect. For commodity prices along I’m not sure if those stats are available on the net. If they are I would love to know about it.

By: marraco

marraco — Thu, 10 Feb 2011 01:35:14 +0000

Wolfram alpha had being completely useless to me. Even simple search like “median income of all the countries of the world” lead to miserable fails.
… or trying to get a commodity price along the time.

By: PlayForFood

PlayForFood — Sun, 06 Feb 2011 14:44:08 +0000

Interesting appoach to knowledge representation by contrasting Alpha’s approach to IBM’s aka HAL. Why not open source the data structure and encourage the worlds hobbist to partake. Thanks for the article, it has me thinking about the pace and direction of US innovation.