Google Knowledge Graph Interview

By   /  May 18, 2012  /  No Comments

Vorhang aufGoogle’s Knowledge Graph has been the subject of lots of attention over the past few days since the announcement. And the focus of a lot of questions, too.

There’s been discussion on chat boards, for instance, about just who’s gotten access and who hasn’t. In a discussion with a representative from Google, The Semantic Web blog has learned that, like many other new Google services, the roll-out is gradual, in order to ensure the system is handling new functions well. First-come, first-served are those who are signed into Google – but then again, not everyone who is signed in. But the plan is to have everyone who’s signed in on board over the next few days, the rep says; so if you are and don’t have it yet, it should be hitting your browser shortly. Those not signed into Google accounts probably have a week or two of a wait left. So far, the rep said that things have been pretty smooth, so Google’s going at the pace it was hoping to.

Of course, many other questions have been asked around what sources are informing the Knowledge Graph. Google in its public blog announcement discussed some of the public sources in which the Graph is rooted – including Freebase, Wikipedia and the CIA World Factbook, which would lead you to conclude some type of Linked Open Data connection. And the representative told us that its 500 million objects also pull from some of the stores Google has built up over the years, including projects it’s undertaken to scan things like books and artwork. So, there are various internal databases of real world objects it’s been indexing, and also external data, he explains.

What about any intersections with, the GoogleBingYahoo!Yandex initative for providing a collection of schemas for structured data markup that focuses on microdata? Sorry – we can’t tell you much about that either. Right now, Google isn’t sharing any more specifics on sources generally or sources in particular. The representative did note that Google hasn’t been talking about more generally trying to derive more objects from around the web for the Knowledge Graph, and that initially the project hasn’t been about open-sourcing what’s out there – though he acknowledged that drawing on a broader set would make sense.

Currently, there is no particular integration with Google Plus, either. This is its own piece of infrastructure, not particularly tied to the social stuff, he notes, though he says he is sure Google will think about that more going forward. Also not in play for now is any design for interaction with or effect on ads. “For now it really is focused on having this database, and search is the first place that it’s popping up.”

Concrete details about future plans, not surprisingly, are not forthcoming, although there is certainly more to come. “The technology is a layer of intelligence that Google can use for potentially [many] things down the line as we go from the world of keywords to understanding real objects,” the rep notes.

Would one of those many things potentially be a Siri-like competitor for Android devices? Again, that’s speculative stuff the rep won’t get into, but why that seems unlikely is because of the Knowledge Graph’s orientation. “I would say the Knowledge Graph is more about knowing about objects and their inter-relationships and attributes, very factual stuff,” he says – such as that something is a bridge and has a certain height and was built in a certain time-frame and it is related to other bridges and it’s in this place that is another entity in the Knowledge Graph. It’s a little less about grammar, and getting at the syntax of a query in order to know what facts to look up to help a user answer it, he explains. “The Knowledge Graph doesn’t address that sort of natural language processing. You could think of that as a different category of problem. It’s not that we are not interested in that problem, but it is not what this launch addresses.”

The response to the Knowledge Graph from the Semantic Web community has generally been positive. “Fundamentally I think it’s great stuff,” says Ivan Herman, Semantic Web Activity Lead at the W3C. While he, like everyone else, would like to have more details on exactly how the Knowledge Graph is doing what it does, the Linked Data alikeness is there. It is “the same thing of referring to a graph of terms and emphasizing the difference between strings and things, whether identified as a URI or with some internal thing…but their Graph is a link among these things. So it’s very much the same and that’s good.”

That said, he also points out that some people have noted that it wouldn’t have hurt Google to at least give a nod to the people and institutions that have worked for years on Linked Data and the Semantic Web. “They could have mentioned that they benefitted from the work a lot of people have done. It’s perfectly okay that they use it, and we’re happy that they are, and it’s okay if they use different terms that go down better for the average user, but [some acknowledgement] would have been nice.”


About the author

Jennifer Zaino is a New York-based freelance writer specializing in business and technology journalism. She has been an executive editor at leading technology publications, including InformationWeek, where she spearheaded an award-winning news section, and Network Computing, where she helped develop online content strategies including review exclusives and analyst reports. Her freelance credentials include being a regular contributor of original content to The Semantic Web Blog; acting as a contributing writer to RFID Journal; and serving as executive editor at the Smart Architect Smart Enterprise Exchange group. Her work also has appeared in publications and on web sites including EdTech (K-12 and Higher Ed), Ingram Micro Channel Advisor, The CMO Site, and Federal Computer Week.

You might also like...

Data Governance 2.0 at the Hub of Data Curation

Read More →