InMotion Ventures portfolio company Wluper is creating a new way to navigate – and your input can help shape it.
Using natural language processing (NLP) and deep learning, the team is building a navigation assistant that understands conversations, not just commands. This enables sophisticated, intuitive journey planning and more contextual travel information.
The team is currently preparing the product for an alpha launch, where early users will be able to test the assistant when using London’s public transport system. Which is where you come in. At training.wluper.com, questions you ask help train Wluper’s NLP models – and make the service more and more effective.
Questions can be short or complex and cover anything transit-related: directions, timetables, price information and more.
“These real-life inputs are valuable because they offer two major insights,” explains Wluper CEO Hami Bahraynian. “One, understanding what the user’s needs are: what information they’re asking for and expecting as a response. Two, the many different variations on questions and the ways they’re asked.
“These help train our neural networks to understand related queries more accurately in the future, even if they’ve never handled the same query before.”
Already, the submitted questions are revealing some interesting patterns.
“A lot of the collected data shows users expect quick responses,” says Hami. “Like ‘Hey, any delays on the District Line that impact me today?’ or ‘Hi Wluper, do you think I’ll catch the next Jubilee to office?’.
“This suggests commuters and travellers mostly know how to get somewhere, yet often need precise additional information regarding their journey.”
Next time you have a question about getting from A to B, head to training.wluper.com. “The Wluper team is thankful for every input,” says Hami.