Register

SpiNNaker’s New Supercomputer Uses AI To Assist With Tough Accents

Tech

photo

Voice-activated virtual assistants don’t understand accents as well as we’d like. SpiNNaker’s new supercomputer could change that.

“I’m sorry, I didn’t understand that.”

This is a common and frustrating refrain from voice recognition tools, and it can be all the more challenging or demoralizing if the speaker has an accent. But a recently switched on supercomputer is aiming, among other things, to reduce the frequency of these misunderstandings over time.

The SpiNNaker supercomputer, activated this past week, mimics the human brain by sending packets of information to a number of destinations at the same time. This mechanism differs from how a standard computer works, sending one packet of information at a time to a single receiver. In their words, the chip-powered computer “breaks the rules followed by traditional supercomputers that rely on deterministic, repeatable communications and reliable computation.”

By mimicking the human brain, the hope is that this computer will be able to assemble received information and learn over time. One of the most immediate and most urgently needed applications of such technology: learning to adapt to a speaker’s voice.

Just as humans can learn to understand individuals with accents over time—certain words will always sound a certain way, and their sound will differ from how it may have originally been programmed—means that virtual assistants like Siri, Alexa, and Cortana could learn and get better at hearing their owners over time. In turn, this means these tools will be effective for more and more prospective users over time. Developers aiming to connect their products to voice assistants can rest easy knowing that the tool will understand their users with increasing ease.

The technology being piloted is twenty years in the making, and yet has a good deal longer to go before it’s consistently effective. For those accustomed to grumbling and then repeating themselves louder or with more enunciation, SpiNNaker’s work can’t go quickly enough. But thus far, it shows promise in creating a voice-assisted world open to people the world over…no matter how they sound.

Learn the latest trends, insights and best practices from the brightest minds in media and technology. Sign up for SMW Insider to watch full-length sessions from official Social Media Week conferences live and on-demand.




Newsletter Subscription

Get the latest insights, trends and best practices from today's leading industry voices.


Watch SMW Live

SMW Insider is a premium video platform that streams more than 180+ hours of talks, presentations, and interviews from leading industry experts.

Subscribe Now

Write for Us

Interested in sharing your ideas and insights with the world? Become a SMW News contributor and reach 300k readers each month.

Apply Here