Keynote: the Future of #smx Research

[messi10]

You're tuned in to the BCI blog where we blog live on SMX West all week. This is the show's keynote, a demo-heavy presentation by Behshad Behzadi , Principal Engineer at Google Zurich. Danny Sullivan with keynote speaker Behshad Behzadi Danny Sullivan with keynote speaker Behshad Behzadi (Photo credit: Search Engine Land) Behzadi is the Director of Conversational Research. Danny Sullivan says Behzadi has done this presentation before at SMX London, and it was a stunning presentation on what's possible with conversational search. Behzadi has been with Google for 10 years, the first seven years working on rankings and the last three years working on future technologies, including Now on Tap.

This is where search is evolving and where Google is strategically investing more. A photo of Captain Kirk is on screen, then a video of Kirk talking to the Star Trek computer. (Audiences absolutely crack up when a 15-second non-skippable YouTube ad plays before the Star Trek clip.) Another music video that Behzadi plays is from the movie "Her", where we see an artificial intelligence operating system. Both of these movies imagine a future where you talk to a machine for answers and help. We are entering this kind of AI experiment, and it will not be in 20 years or in 200 years. That's the direction we're headed in now, and as we'll see in its demos, we're already fax number list pretty close. How research has evolved to this point Research timeline slide The evolution of research over time An early breakthrough in how search functions understood synonyms from 2002.

It shows a list of queries that contain "cs" and how Google might interpret "cs" differently: “cs doctorate admission in california” —> cs = computer science (Google interpretation) « daily price for cs equity funds » —> cs = credit suisse “cs bank hayfield” —> cs = state of citizens "2007 CS World Cup" —> CS = counter-strike "bus amsterdam cs to airport" -> Google interpreted cs = central station Another big step forward was universal search in 2007. The whole problem of results becomes much more difficult because we are comparing apples and oranges. Then, in 2012, the Knowledge Graph and Google's understanding of "things, not strings" was the next step forward in understanding the real world. 2 billion entities, 54 billion facts, 38,000 entity types — and growing. (Behzadi updates these numbers each time he presents.)