Weotta

Jun 10, 2014

Follow CommentsFollowing CommentsUnfollow Comments The vast possibilities of things to do on a free night out can be as overwhelming as trying to pick a good movie when there are hundreds of options. Netflix Netflix is tackling the movie problem with algorithms that learn what kind of Hollywood fare you enjoy. Weotta is an app that not only learns what you like, but better understands the context of what you’d like to do in the moment — better, it claims, than Google Google . The San-Francisco-based local local -discovery service updated its app on Tuesday with a series of enhancements to make it smarter than before. It can now understand “soft” queries that Google’s advanced search engine might not fully process, according to co-founder and CEO Grant Wernick. “Google does a good job at text search,” Wernick says. “They’re the best.” But it still can offer up a long list of largely irrelevant answers to a question like “What local events are happening next Thursday?” “Google would have to understand context, date range and events,” says Wernick. Currently Google does offer a ribbon of local results if you search for something like “sushi restaurants in San Francisco.” But Weotta wants to push the envelope by answering more detailed questions like “What’s a good sushi restaurant where I can bring my kids this weekend?” It’s “the kind of queries you’d ask a friend,” says Wernick. “The smarter search technology allow you to search a much wider range as opposed to looking at a list.” “We’re searching for deeper level of features that are part of the data we found,” says Jake Perkins, Weotta’s CTO and co-founder. ”That allows us to do semantic search so we know about what you mean, which may not be the exact phrases people use when describing a place. Like synonyms.” When it comes to serving local searches, Google is in a sense a victim of its own success, having crawled so much of the web that processing specifics questions about local, contextual events becomes incredibly complicated. “They have to serve everyone,” says Perkins. “They’ve got so many websites and so much data, but then they may not actually have the events and the times they need to answer that question.” Weotta’s language enhancements come at a time when the natural-language processing (NLP) technology that Nuance and Google are developing is only getting better. It means consumers are likely to get more and more accustomed to wording queries they make to computers and digital personal assistants like Siri, in the same way they might to a human. That’ll certainly be the case over the coming years. Weotta, incidentally, uses a stack interface that’s reminiscent of Tinder; you swipe through the images of suggested local events and swipe down if you’re interested. Users end up getting mesmerized by the process, Wernick says, which Weotta claims to have invented and published before Tinder did.