Google: Our understanding of search questions 'pretty limited'

Search engine operator admits understanding of search queries and existing data available "pretty darn limited", but it's making "baby steps" toward vision of providing best answers to search questions as fast as possible, says exec.

Google's aim for its search service is to give users the best answers to their questions as quickly as possible, but while it is moving toward this goal, a Google executive said it is a "long road ahead" before this vision can be realized.

Amit Singhal, senior vice president and Google Fellow, said the company's ability to deliver such an experience depends on it understanding users' questions as well as "truly understanding all the data that's out there". However, he noted that "its understanding is pretty darn limited" currently.

"Ask us for the '10 deepest lakes in the U.S.' and we'll give you decent results based on those keywords, but not necessarily because we understand what depth is or what a lake is," he wrote in his Google+ account on Thursday.

To this end, he pointed out that Google had bought Freebase, an open source knowledge graph, in 2010 to use it as a tool to "aid the creation of more knowledge".

With Freebase's 12 million canonical entities as a starting point, Google is "building a huge, in-house understanding of what an entity is and a repository of what entities are in the world and what should you know about those entities", Singhal explained in a separate interview with tech news site Mashable in February.

"A knowledge base is huge compared to the word index [which is what today's search algorithms are based on] and far more refined or advanced," he stated in the interview.

To date, Google has increased the 12 million entities to over 200 million, the executive noted. This development, however, represents just "baby steps" toward the "virtuous cycle" of creating more knowledge and insight to the world's existing data and there's a "long road ahead", he said.

Singhal's Google+ post came after the Wall Street Journal reported that the Internet giant is working to completely revamp its search function to make it look more like "how humans understand the world".

Some major changes will show up in the coming months, according to unnamed sources in the WSJ report, but Singhal stated that Google is undergoing a years-long process in transiting to "the next generation of search".

Google isn't the only one looking to create a semantic search engine. Yahoo, too, told ZDNet Asia in a 2010 interview it was looking to "unravel the mystery of understanding the world in semantic terms". Raghu Ramakrishnan, chief scientist of audience and cloud computing research at Yahoo, said then that this is a great challenge that will not be solved in the next 5 to 10 years.


You have been successfully signed up. To sign up for more newsletters or to manage your account, visit the Newsletter Subscription Center.
See All
See All