The result is that those providers, like Google, can not simply hope to coax faster performance out of their systems, they have to rethink how their processes identify, manage and provide whatever the customer is seeking. And that requires reimagining the way the brain interacts with technology. JL
Matt McFarland reports in the Washington Post:
Google is shifting toward a more natural way for users to get their questions answered. On the app, there’s no need to type due to voice recognition. And a user won’t always have to click to another Web site to find an answer.
Google has updated its search app to understand more complicated questions so that it can give precise answers rather than a list of Web sites that likely include the answer.
The search engine now includes a feature that parses the sentence structure of questions, then uses its knowledge to provide an answer. Once Google is confident what the user is interested in, it can look at its bank of knowledge to find the answer. This information is stored in its Knowledge Graph, which it unveiled in 2012. It includes lists of presidents, sports rosters, census data and more.
But there are limits to how much Google can directly answer. The feature is designed to work on superlatives, specific time points and some complex combinations of the two. For example, it can tell you who the tallest players on the Golden State Warriors are, with “the tallest” being a clue of the query’s intent. Other examples include, what the population of Japan was in 1990, or who was president when the Chicago Cubs won the World Series. (It’s been a while.)
Google receives many queries about celebrities, as well as health and sports trivia. The app’s new function is built to excel at questions like “What songs did Taylor Swift record in 2014?”
Some complex questions will still confuse it. For example, asking “Who was Dakota Johnson’s mom in the movie?” will trigger movies that Johnson’s real-life mother was in, not the name of the actress who played Johnson’s mother in “50 Shades of Grey.” While the technology illustrates an improvement in understanding a user’s intent, it’s not perfect.
Google is shifting toward a more natural way for users to get their questions answered. On the app, there’s no need to type due to voice recognition. And a user won’t always have to click to another Web site to find an answer.
These new capabilities aren’t relying on deep learning, a type of artificial intelligence that’s gained a lot of buzz lately as computers drive cars and beat video games. But we can expect Google to continue to add more human-like intelligence to search as it looks to stay ahead of competitors.
0 comments:
Post a Comment