updated 11:55 pm EDT, Wed March 14, 2012
Engine to provide direct answers to queries
Google is reportedly preparing to revamp its search engine, embracing "semantic search" technology to improve relevancy. The algorithms will be adapted and expanded in an attempt to understand the meaning of the words in a search query, allowing the servers to respond with a direct answer to certain questions.
Google executive Amit Singhal suggests the company has been building a library that correlates millions of items such as places, people and things, according to excerpts from an interview (sub. required) posted by the Wall Street Journal. The semantic search technology will draw from the database as it associates different terms.
Singhal describes the semantic system as akin to "how humans understand the world," rather than relying solely on existing web pages to provide relevant answers. The company already offers a limited number of direct responses, such as answers to math questions and currency conversions, however the revamped system will provide a deeper level of data related to search terms.
The semantics search technology appears to follow the path that Apple took when developing its competing Siri "intelligent personal assistant." Recent leaks suggest Google's alternative, known as Assistant, will provide a "personalization layer" that pairs a voice recognition engine with algorithms that understand relationships between different types of data.
Google is expected to launch the first semantic search capabilities sometime in the next few months, though exact dates have yet to be confirmed.