Wednesday, January 07, 2015

Search: The Final Frontier

Search as we know it now is based on words and text, but this is limiting in many ways.  Merely searching the web of links and URLs no longer works in the sophisticated evolution of the Web.   How do we push out of those traditional boundaries of search?   Bing Director of Search at Microsoft's Bing Stefan Weitz has come out with an excellent book called simply Search predicting the world of search engines in the future, particularly when "every device, every object that surrounds us, and every person is connected and we have systems that can distinguish patterns from the noise."

Of course, when we think of  search, it is language-based.  The problem of our current search is the "low resolution" approach of describing things.  The statement a picture is worth a thousand words in this aspect is really apt: it literally takes at least a thousand words to precisely to define the attributes of even the most basic object as a table (such as dimensions, materials, country of origin, date of purchase, etc).  This is for a simple object, too!

According to Weitz, searching and describing the world required a method that both search engines and humans could process: words mapped to pages, and those same pages mapped to information.  Even audio-visual information are associated through text.  

What if search is described by very fine details not by human language but by things?  What if  sensors through a billion connected devices become an Internet representation of a physical object?  No longer do we have to rely on language as a universal descriptor, but suddenly the world can be modeled in seemingly endless detail using facets that have little relation to language.

The boundaries between digital and physical worlds blend into each other and complement one another.   In this new search world, the increasing digitization of the world means that everything can be described at higher resolutions than ever before.

Timing is everything, too, as the Web has moved away from being an information repository to a bridge that transports data and enables services to happen in real life.  Weitz likens this transformation of search from a large library into a digital proxy for the physical world.   Just as humans are unable to think about things isolation, the Web needs to perceive the world holistically, away from a Web of words to a Web of the world.  When search understands every characteristic of every object on the planet, it will be finally able to make those connections.  I've written about this in previous posts about the current evolution of the technology and how it shapes our thinking as information professionals.

These connections are called graphs.  By linking these disparate graphs, the commonalities that emerges will assemble a complete picture of the world and everything in it.  Every person and object can be described in hundreds of ways, from photos from Facebook to your personal weight when you're on a scale, to the energy consumed by your smartphone.  In this smart-world, machines build models that re-create reality, becoming aware and sensitive to our needs and inquiries.

Such machine learning - an area of artificial intelligence (AI) - allows search systems the capability to make sense of the world.  Whether we call this brave new world a Web of Things, Web 3.0, smart web, or intelligent web, is still to be determined.  But the potential for for the interconnectedness of this system that will change the way we search is almost as exciting as when the Web first became ubiquitous.  The future of search feels bright indeed.  



No comments: