Monday, December 27, 2004

The Search Engine Of The Future

The New York Times is reporting that IBM has come up with a search engine that makes use of semantics for doing its searches.

Yahoo is an example of the first generation search engines. It used simple keywords to do its searches. Second generation engines like Google and Ask Jeeves used some processing logic to reply to searches. These engines have been great at searching structured random data. What has been missing is the discovery of information. The human mind can scan data and implicitly deduce relations in the content. This discovery process has been sorely lacking in current generation search engines.

IBM's search engine goes one step further. It can deduce relations based on context. For example, after scanning a news article about Canadian politics, IBM's system responded correctly to the question, "Who is Canada's prime minister?" even though those exact words didn't appear in the article. IBM's search engine searches across multiple languages and provides translations. It can thus create context in across languages and random data.

Google and Yahoo are working hard on providing ability to search in non text data like audio and video. Those search capabilities combined with the semantic search
unveiled by IBM will form the future of searching.

Pair this search with one more phenomenon making the rounds - Social Networks. These depend on the so-called FOAF - Friend of a Friend. The idea is to build a web of trust. The trust is strongest between two acquaintances and diminishes with distance. This idea can be extended to identify trustworthiness of documents. If a document from a trusted person refers to another, then that too can be considered trustworthy and so on down the line.

With these tools linked together and fully implemented, one could ask a theoretical question, say, "Is company X worth investing?" The 3G search engine will go and trawl through all available information building context. The FOAF filter will be applied to the information to cull the useful data out of it. Finally, an accounting and financial rules AI engine can be applied to the data to come up with an answer. With the incredible leaps in CPU processor speed and the cheap disk space available, this could soon be a reality for consumers.

At I.B.M., That Google Thing Is So Yesterday


Blogger Ryan Melvin said...

I think it would be great if IBM got that to work. I have often wished that google could understand what I was asking when I searched for something. Even with its "second-generation" technology, google often still leaves me to search through a sea of information which may or may not be relevant to what I want to know. Maybe IBM and google could get together.... With IBM's technology and Google's massive database, they would be an internet superpower.


February 4, 2005 at 3:43 PM  

Post a Comment

<< Home