The 2009 Semantic Technology Conference kicked off yesterday. Between now and Thursday, there’ll be a total geek-out covering all that’s coming in Web 3.0 (whatever that actually means!).
Of special interest to me is the semantic search day. It includes a fabulous keynote panel featuring some of the top figures in semantic search research. I’m particularly looking forward to hearing Andrew Tomkins of Yahoo and Peter Norvig of Google. I’ve been following their work for a long time, and they are top-class thinkers. Also on the panel will be Scott Prevost of the Powerset division of Microsoft’s Bing, Tomasz Imielinski of Ask, Riza Berkan of Hakia, and William Tunstall-Pedoe of True Knowledge.
I spend a lot of time researching information retrieval on the Web and how it may affect search marketing’s future. A frequent question I get on my travels is: “What is the semantic Web?”
Mostly, people seem to think there’s another Web being built somewhere, getting ready for launch. That’s not the case. It’s more about bringing meaning to the Web we already have. Now you’re scratching your head and thinking, “I know what Web pages mean.” And you’re probably right. Point is, computers don’t. Computers can figure out syntax, but that doesn’t say they understand meaning.
There’s a lot of research going on around sentiment analysis, as in, “I know people are talking about my brand, but are they saying nice things?” Let’s face it, we all know there are two ways you can think about the word “bad.” Sometimes it means good. And, not so mysteriously, sometimes it can mean not very good at all.
You can take that to another level. In New York, many tourists buy “I Love New York” T-shirts. However, the word “love” is replaced by a heart. You and I know what that means. But how would a computer?
The Web as we now know it was conceived more as a library in which anyone and everyone is able to contribute. It’s built on the notion of being able to gather the world’s knowledge and have it in one place. But there are no real rules.
And when it comes to search, the Web is essentially syntactic. This is why information retrieval on the Web is such a fascinating science. If you know the name of something, it’s typically not too difficult to find it on the Web using a search engine. But when you don’t know the name of something and have to use a description, search usually falls over.
Tell any New Yorker you had a sandwich made with rye bread, corned beef, Swiss cheese, sauerkraut, and Russian dressing but can’t remember what it’s called. He’ll tell you it’s a Reuben. But just throwing a few ingredients into a search engine may not get you such a quick or even correct response.
This is where the semantic Web comes in. It’s more about Web services, where machines can work together to perform inferences in the way people do. The idea behind the semantic Web is to try to turn information on the Web into something with a much more clearly defined meaning.
At this time, maybe the earliest attempts involve using XML to embed structured data into a document alongside unstructured text. And here’s where I opt out of too much talk about the underlying technology for fear of you falling asleep.
Let’s go back to meaning. In most languages, syntax is how you say something. Semantics is about the meaning of what you said. So, even though the Web as we know it was developed to allow all computers to talk to each other, these computers don’t actually know the meaning of what they’re talking about.
The semantic Web isn’t about artificial intelligence, with computers learning how to understand human language. We’re talking about a concept whereby computers will have enough semantics to allow them to solve well-defined problems through the sequential processing of operations.
It may be that a software agent doesn’t even come close to the conclusions that a human is capable of. But it may contribute to building a better Web than the one we have today.
And all I can say after that is: I hope you get my meaning!
Join us for a one-day Online Marketing Summit in a city near you from May 5, 2009, to July 1, 2009. Choose from one of 16 events designed to help interactive marketers do their jobs more effectively. All sessions are new this year and cover such topics as social media, e-mail marketing, search, and integrated marketing.