(Annette Moser-Wellman) We each have our own tricks for searching the Internet for the information we need. I've found that if I drop the "s" off of plural words in Google, I get what I'm looking for faster. But the day is coming, and maybe just around the corner, when search will rapidly improve and trump our current basket of tricks. Further, the same technologies that will streamline search will have big implications for online advertising.

The promise of better search is being driven by smart, or semantic technologies, that extract meaning from information. At the forefront of the artificial intelligence movement, semantic technologies are revolutionizing the way in which we access data on the Internet. I interviewed Matt Colledge, Chief Executive Officer of Idilia, a software firm that researches and develops applications for semantic processing. For the last eight years, Matt's firm has been pioneering at the edge of artificial intelligence.

Idilia's breakthroughs are in the area of "word sense disambiguation" - the capability to automatically resolve the meaning of words in context. It sounds scary but bear with me and you'll catch of glimpse of the radical changes coming.

"Resolving words in context is something that human beings address quite naturally. In fact, we reuse words to mean many different things, and when we read or speak, we have no trouble resolving which meaning is being used based on the context," Matt explains. "There isn't any software today that allows a computer to do that same thing.

"A computer that's reading text sees every word as any word that is spelled the same way. In fact, the computer is unable to differentiate the text J-A-V-A to mean a dance, an island in Indonesia or a programming language. This problem is generally classed as an artificial intelligence type of problem," he said. "We set out to develop some software that would allow a computer to detect and correctly assess the meaning of a word at a level of accuracy that was comparable, if not equal to, that of a person."

It's not hard to imagine how this will change search. When you want to book a flight to New York, the engine doesn't currently know whether you're talking about book in the sense of 'reserve' or book in the sense of a book that you read. But this is what word sense disambiguation does. It acts as a person does with naturally-expressed language – it places the words in context. So rather than trying explain what you are looking for, you'll be able to query information on the Internet with much greater accuracy and it will automatically "understand" your meaning.

"Our technology allows for the search engine to know that you're talking about booking an airline flight. Through the power of tagging, the software figures out that this is the sense meaning 'make a reservation' and not the sense meaning 'a piece of literature.' So now when you start looking for information, the computer can start to act like a librarian would. It can start to interact intelligently with you. When it's not sure of your meaning, it can actually ask you, 'Were you looking for this particular meaning?' And this technology may well operate with more accuracy than a librarian," he said, noting that in accuracy tests, Idilia's software performed within 5% to 10% of a human judgment rate.

So what are the implications for media companies of technology like this? First, think advertising: word sense disambiguation will have the capability to match advertising to content with surprising precision.

Matt explains: "This technology makes it possible to do much more pervasive matching or much more comprehensive matching. It creates higher recall - matching more ads to more content and more accurately than the technology out there today. In our experiments, we often see an increase of 100% in the number of relevant ads you can match to queries. So, fundamentally, you're changing the technology that sits underneath the hood of all of these search applications."

An additional benefit: once you understand the meaning of a word, you can understand whether that word has a negative connotation or derogatory connotation -- an issue that has been traditionally difficult when trying to match advertising to content. There was a famous example of a luggage ad being matched against a document that talked about a body being discovered in a suitcase. Clearly, these are classic mistakes of automation that word sense disambiguation is able to tackle.

But I suppose what makes this technology so thrilling is the ability to free up access to information on the Internet and make it open to wide-ranging inquiry.

What did it take to reach this break-through? "There are almost 30 people at Idilia who are specialized in computation linguistics, machine learning, software engineering and pure linguistics, and it took a team of people eight years to construct this technology and make this giant integrated system work," he said.

When I asked if eight years of development time was a painful process, he replied, "Well, it's painful if you don't like the problem. It's quite exciting if you find the problem interesting."


What do you think? Please share your thoughts, experiences and reactions by clicking on the comment button below.

Annette Moser-Wellman is President of Firemark, Inc., an innovation consultancy, and author of Six Competencies of the Next Generation News Organization and Running While The Earth Shakes: Creating An Innovation Strategy To Win In The Digital Age, both published by the Media Management Center. She teaches in MMC's Advanced Executive Program and Digital Strategies for Media Executives seminar.

This TechScout article is part of a series of Moser-Wellman interviews commissioned by the Media Management Center to explore opportunities and insights at the intersection of technology and the news media. Click here to view other articles in the TechScout series.