Google search algorithm improvements will help it understand complicated queries and provide better answers.
20 Great Ideas To Steal In 2013
(click image for larger view)
Google marked its fifteenth anniversary on Thursday with a revision its search engine algorithm called "Hummingbird."
The company introduced Hummingbird at Google senior VP Susan Wojcicki's old Menlo Park, Calif., house, where Google in its early years operated out of the garage.
Google updates its search algorithm frequently, on the order of several hundred times a year. The changes produced by most of these adjustments tend to be too subtle to notice.
But the company gives names to its major architectural updates, like Panda (February 2011), an effort to reduce the prominence of low-quality content in search results, and Caffeine (June 2010), a rewrite of the company's Web indexing system so that Google could provide faster search results.
Google has also given names to content-oriented search changes, like Search Plus Your World and Universal Search.
The Hummingbird update expands Google's use of its Knowledge Graph, introduced last year as a way to help its search engine understand the relationships between concepts rather than simply matching keywords in documents. The Knowledge Graph structures data, so that a search for, say, Marie Curie, returns facts about her contributions to science, her life, her family and other related information, not all of which are necessarily contained in the same document.
Hummingbird expands Google's use of the Knowledge Graph so that its search engine can provide answers to queries that don't necessarily have simple answers. In a blog post, Amit Singhal, senior VP of Google Search, points to a search like "Tell me about Impressionist artists," which now returns a broad set of appropriate facts when submitted through a mobile device.
The Knowledge Graph also helps Google understand when a follow-up search makes reference to a previous search. For example, if you ask the Google Search app for "pictures of the Washington Monument" and then ask, "How tall is it?", Google will understand that you're referring to the Washington Monument instead of treating your query as a separate question.
The search update also adds a comparison tool. For example, the query "Compare butter with olive oil" returns an organized set of data covering nutritional information. This is the sort of query that specialized search service Wolfram Alpha handles well.
Google Search on mobile devices has been redesigned and in the coming weeks, updates to the iOS and iPad versions of the Google Search app will support Google Now notifications across multiple devices. Singhal said that if you tell your Nexus 7, "OK Google. Remind me to buy olive oil at Safeway," and you visit Safeway with your iPhone, you will still receive the Google Now reminder you set.
Looking ahead, Google can be expected to continue pushing the development of predictive search (Google Now) and voice search, because typing on a mobile device is slow and often impractical.
We welcome your comments on this topic on our social media channels, or [contact us directly] with questions about the site.