Wluper understands better: take the quiz and see if voice assistants can handle common transport questions

May 29, 2018 · London-based Wluper builds conversational assistant technology for transport and navigation

536294 large

Intelligent assistants like Alexa and Siri can already handle a lot of common questions. But they’re not always so helpful when it comes to transportation questions. When we’re on the go, we want to know the fastest way to get from A to B and the best coffee we can pick up along the way. This is where London-based Natural Language Processing (NLP) startup Wluper comes in.

Backed by Jaguar Land Rover’s InMotion Ventures, Wluper is enabling conversational interfaces for the transportation and navigation space by solving a specific pain point: the poor navigational intelligence.

To illustrate the difference, Wluper has created quiz.wluper.com, where you can assess a series of common transportation questions and decide whether they can be handled by machines or not. As well as showing you how you did, the quiz answers demonstrate how Wluper’s NLP and logic engine perform against generic systems like Alexa and Siri.

"The tech giants have the goal of create a voiceover for our lives, which might take another 15 or 20 years," says Wluper Co-founder Hami Bahraynian. "We need intelligent agents now, created for a well-defined purpose."

Wluper is building domain-specific NLP using differentiable end-to-end Deep Learning models to create the ultimate transport assistant technology. "One that not only understands humans, but understands like a human," says Hami Bahraynian. "It’s a big difference."

Existing digital assistants run into trouble due to their goal of creating general-purpose Artificial Intelligence; the team around Wluper is narrowing down the problem and tackling a specific domain transportation - in order to provide a better user experience.

The startup’s first Machine Learning models are outperforming industry benchmarks already. In an alpha test earlier this year, which covered London’s public transport, Wluper was even able to beat Amazon’s Alexa and Apple’s Siri in some use cases, specifically in understanding human-like transport- related questions, rather than one-way commands.

"Since we can make clear assumptions about what the user wants - directions - we can understand more complicated and very naturally asked questions," says Hami Bahraynian. "Wluper can handle multi-intent queries, and ultimately follow- up questions, enabling something much more like a true conversation."

Earlier this year, Wluper was selected for the NVIDIA Inception Program, a virtual accelerator program that helps AI startups during critical stages of product development, prototyping and deployment, by providing Deep Learning expertise and best-in-class training.

The Wluper team sees a huge number of applications in the transportation space for the technology it’s building, including automotive, GPS navigation systems, ridesharing services and public transportation apps. There is also enormous potential for it in autonomous vehicles.

"How are next-generation Uber users supposed to interact with driverless cars? How will they navigate through cities?" says Hami Bahraynian. "A new interface is needed for the user to ask the car to stop at the next corner or change the route more intuitively."

Wluper, based in London, is backed by Jaguar Land Rover’s InMotion Ventures and is currently raising a Seed round.

Read Complete Article

RT @LDN_talk: Are you using online government or public services, e.g. to pay your council tax, make a complaint or register to vote? Tell…

I was honoured to speak tonight at the opening of @SomaliWeekFest at the @britishlibrary . Somali Londoners are one… https://t.co/2rvvikVaDC

#TechX365: As more and more people work out of coffee shops on public WiFi, greater caution needs to be taken that… https://t.co/Z0HqKQ3o05

Impressive video and good explanation https://t.co/MVgcq4HNY7

'I'm sorry, Dave': 9 real times AI has given us the creeps https://t.co/RUjVVb0OVu by @gcnichs