Comments on: Google's search challenge: Making computers think like humans

Google has untold thousands of computers, but the head of the company's search work says the difficulty is making them understand the human mind.

[dargon19888]

Getting a computer to think like a human?

You first have to understand how a human thinks.
If you can tell me how my wife or any woman for that matter thinks, you've got yourself a Nobel Prize!

[Fegna]

Udi say that the computer has to think like human, not think like women. You know that are different ideas, or no? :-)

[benjamin straight]

What a challenge!

[SenorFrog]

"Hal? Open the page, Hal". "I'm sorry Dave, but I don't think you really want to go there. I know what you're really looking for when type in a search for Japanese Manga".

[The_Decider]

You are not trying to get a computer to think like a human, which is impossible, since computers do not think. You are talking about a computer being able to figure out the intent of what someone typed. That is impossible as well. It makes a good story and certainly enthralls ignorant journalists, but has no basis in reality.

[mtoc]

how about making computers "think" like computers! living organisms THINK! computers
CALCULATE! too much effort has been made to complicate programs rather than to ease the execution of same. which is why VISTA is a pain. ergonomics, display, process
speed, simplification etc. would be a better goal. can you imagine running your car like you do
you PC!

[expert_user]

Wow, talk about a poor piece of software. After trying it, I couldn't even uninstall. Don't try this on your machine

[ralfthedog]

On the positive side the spammer now has all of your credit card numbers.

.

Never install software pushed by a spammer. Never go to a website pushed by a spammer.

.

PS. Please enjoy all the pop up adds.

[Crunchy Doodle]

In 1978 my wife and I worked on a project called Datavault. The idea of Datavault was to be able to make query searches of large textual unindexed databases. I helped build the custom nine CPU computer that would scan through the text and develop a "likeness factor" of the data. When a query was made, these likeness factors were matched to the query and presented to the user. Does any of this sound familar? The downfall of this project was the same problem faced by Google engineers today - making the result relavent for a human. I remember making the query for "Banks by a river"; meaning financial institutions located near a body of water, and getting back something about the edge of the Mississippi. I guess the more things change, the more they stay the same.

[jaken37]

Skynet

[scottjarus]

"We have to understand as much as we can user intent and give them the answer they need," Manber said.

Cognition Technologies' extensive semantic map of the English language and its vast understanding of synonymy, disambiguation within context, and semantic structure are key components to Search engines evolving toward Manber's goal of "human understanding and intent". The first step in this evolutionary process is "understanding" the meaning of a user's search query and then matching them to the most relevant results based upon an "understanding" of the content being searched. These technological early steps toward an intelligence understanding of user intent are available now by employing Cognition 's Semantic NLP technology. See for yourself at www.cognition.com.

Scott Jarus, CEO, Cognition Technologies