Last week I visited the Thomas J. Watson Research Center in Hawthorne, 20 miles north of New York, to hear six I.B.M. researchers describe their company’s concept of “the future of search.” Concepts and demos are different from products being shipped and sold, so it is unfair to compare what I.B.M. is promising with what others are doing now. Still, the promise seems great.
Two weeks before our meeting, I.B.M. released OmniFind, the first program to take advantage of its new strategy for solving search problems. This approach, which it calls unstructured information management architecture, or UIMA, will, according to I.B.M., lead to a third generation in the ability to retrieve computerized data. The first generation, according to this scheme, is simple keyword match - finding all documents that contain a certain name or address. This is all most desktop search systems can do - or need to do, because you’re mainly looking for an e-mail message or memorandum you already know is there. The next generation is the Web-based search now best performed by Google, which uses keywords and many other indicators to match a query to a list of sites.
I.B.M. says that its tools will make possible a further search approach, that of “discovery systems” that will extract the underlying meaning from stored material no matter how it is structured (databases, e-mail files, audio recordings, pictures or video files) or even what language it is in. The specific means for doing so involve steps that will raise suspicions among many computer veterans. These include “natural language processing,” computerized translation of foreign languages and other efforts that have broken the hearts of artificial-intelligence researchers through the years. But the combination of ever-faster computers and ever-evolving programming allowed the systems I saw to succeed at tasks that have beaten their predecessors.
One example is question answering. Google-type search engines are fabulous at retrieving random data, but mediocre at handling subtler queries. Using Google or Ask Jeeves, you can eventually find out how many of the world’s Web pages are in each of the major languages, but it’s slow and frustrating compared with finding out, say, Mozart’s birthplace. Jennifer Chu-Carroll of I.B.M. demonstrated a system called Piquant, which analyzed the semantic structure of a passage and therefore exposed “knowledge” that wasn’t explicitly there. After scanning a news article about Canadian politics, the system responded correctly to the question, “Who is Canada’s prime minister?” even though those exact words didn’t appear in the article.
The Semantic Analysis Workbench, demonstrated by Eric Brown and Dave Ferrucci, showed another way of exposing latent meaning. The I.B.M. officials said the best use for this technology would be customer-support call centers: As representatives took notes on the problems people were having with their cars or computers or prescription drugs, automatic interpretation of the results would reveal useful patterns. Arthur Ciccolo, an I.B.M. strategist for its unstructured-information project, said that call centers would be the first place for new search systems to be applied. Genomic-research projects, where unexpected correlations can be crucial, might be the second. But the demonstration suggested another likely market, since every bit of sample text was a transcript of intercepted phone calls, apparently among people suspected of terrorism. (”He made two calls from Frankfurt on these dates … “) Whether these were real, I still don’t know.
Salim Roukos demonstrated a system I would like to have tomorrow: an assortment of news headlines, roughly comparable to Google News, but from non-English language sources. The system automatically - and comprehensibly - translated the headlines and leads of each article. If you wanted to read more, you pressed a button and in 15 or 20 seconds had a good-enough translation.
Mr. Ciccolo, the search strategist, said that in a way his team was trying to match - and reverse - what Google has achieved. “As Google use became widespread, people began asking why it was so much easier to find material on the external Web than it was on their own computers or in their company’s Web sites,” he said. “Google sets a very high standard for that Web. We would like to set the next standard, so that people will find it so easy to do things at work that they’ll wonder why they can’t do them on the Internet.” How soon might this happen? He said, with a chuckle, “Well, if I could freeze what everyone else is doing, it could be in two years.” The great part is, the competition won’t be frozen. At least this part of the future looks bright.
J’aimerais bien voir les résultat de leur système de traduction automotique de titres de journaux….
