Technology: Google details AI work behind Project Euphonia's more inclusive speech recognition - PressFrom - US
  •   
  •   
  •   

TechnologyGoogle details AI work behind Project Euphonia's more inclusive speech recognition

02:40  14 august  2019
02:40  14 august  2019 Source:   techcrunch.com

Google Calendar is back after global outage

Google Calendar is back after global outage Some users got 404 messages instead of meeting and event information. Problems started to crop up early in the morning, and Google tweeted that it was "experiencing a service disruption." Google didn't immediately have a comment on what caused the problem. A message, posted to the G Suite Status Dashboard at 7:22 a.m. PT, said the company was investigating reports of issues with Google Calendar and that affected users were unable to access the service. Google said that "system reliability is a top priority." Users reported trouble opening Google Calendar on desktop as well as mobile devices.

Google ’ s Project Euphonia helps make speech tech more accessible to people with disabilities. This is intended to remedy the problem of AI bias created by limited training data. Because speech software like Google Assistant is built to respond to the majority of voices, it doesn’t work for those in

Google is researching ways to improve speech recognition including for people with speech impairments, partnering with nonprofits and volunteers. To help solve this problem, the Project Euphonia team—part of our AI for Social Good program—is using AI to improve computers’ abilities

As part of new efforts towards accessibility, Google announced Project Euphonia at I/O in May: An attempt to make speech recognition capable of understanding people with non-standard speaking voices or impediments. The company has just published a post and its paper explaining some of the AI work enabling the new capability.

The problem is simple to observe: The speaking voices of those with motor impairments, such as those produced by degenerative diseases like amyotrophic lateral sclerosis (ALS), simply are not understood by existing natural language processing systems.

You can see it in action in the following video of Google research scientist Dimitri Kanevsky, who himself has impaired speech, attempting to interact with one of the company's own products (and eventually doing so with the help of related work Parrotron):

Apple, Google continue inclusive push with new emojis

Apple, Google continue inclusive push with new emojis Apple and Google are rolling out dozens of new emojis that of course include cute critters, but also expand the number of images of human diversity. The announcement coincides with Wednesday's World Emoji Day . Apple Inc. is releasing new variants of its holding hands emoji that allow people to pick any combination of skin tone and gender, 75 possible combinations in all. There are also wheelchairs, prosthetic arms and legs, as well as a new guide dog and an ear with a hearing aid. And then there's the sloth, the flamingo, the skunk, the orangutan, as well as a new yawning emoji. New emojis routinely pop up every year.

Today at Google I/O, Google unveiled Project Euphonia , to explore how artificial intelligence can better recognize those with Voice recognition technology doesn’t work today for people with speech impairments because no one has But in order to get there, Google will need more training data.

Google Cloud Speech -to-Text enables developers to convert audio to text by applying powerful neural network models in an easy-to-use API. Google supports more than 10x proper nouns compared to the number of words in the entire Oxford English Dictionary.

The research team describes it as following:

ASR [automatic speech recognition] systems are most often trained from 'typical' speech, which means that underrepresented groups, such as those with speech impairments or heavy accents, don't experience the same degree of utility.

...Current state-of-the-art ASR models can yield high word error rates (WER) for speakers with only a moderate speech impairment from ALS, effectively barring access to ASR reliant technologies.

It's notable that they at least partly blame the training set. That's one of those implicit biases we find in AI models that can lead to high error rates in other places, like facial recognition or even noticing that a person is present. While failing to include major groups like people with dark skin isn't a mistake comparable in scale to building a system not inclusive of those with impacted speech, they can both be addressed by more inclusive source data.

Google ad shows off the Pixel 4’s new air gestures and face login

Google ad shows off the Pixel 4’s new air gestures and face login If you can't beat 'em, join 'em. Google seems to be following this old adage of late, showing off the Pixel 4's design early

Project description. Library for performing speech recognition , with support for several engines and APIs Quickstart: pip install SpeechRecognition . See the “Installing” section for more details . Note that the versions available in most package repositories are outdated and will not work with the

Google has demonstrated Project Euphonia , an effort under its AI for Social Good program that utilizes artificial intelligence to improve speech recognition technology. Google partnered with the ALS Residence Initiative and ALS Therapy Development Institute non-profit organizations to improve its

For Google's researchers, that meant collecting dozens of hours of spoken audio from people with ALS. As you might expect, each person is affected differently by their condition, so accommodating the effects of the disease is not the same process as accommodating, say, a merely uncommon accent.

A standard voice-recognition model was used as a baseline, then tweaked in a few experimental ways, training it on the new audio. This alone reduced word error rates drastically, and did so with relatively little change to the original model, meaning there's less need for heavy computation when adjusting to a new voice.

The researchers found that the model, when it is still confused by a given phoneme (that's an individual speech sound like an e or f), has two kinds of errors. First, there's the fact that it doesn't recognize the phoneme for what was intended, and thus not recognizing the word. And second, the model has to guess at what phoneme the speaker did intend, and might choose the wrong one in cases where two or more words sound roughly similar.

Apple joins Google, Facebook, and Microsoft in data-sharing project

Apple joins Google, Facebook, and Microsoft in data-sharing project The Data Transfer Project wants to make it easy to move data between services

Through Project Euphonia , Google partnered with ALS Therapy Development Institute (ALS TDI) and ALS Residence Initiative (ALSRI). So, Google set out to record thousands of voice samples. One volunteer, Dimitri Kanevsky, a speech researcher at Google who learned English after becoming

Google has demonstrated Project Euphonia , an effort under its AI for Social Good program that utilizes artificial intelligence to improve speech recognition technology. Google partnered with the ALS Residence Initiative and ALS Therapy Development Institute non-profit organizations to improve its

The second error in particular is one that can be handled intelligently. Perhaps you say "I'm going back inside the house," and the system fails to recognize the "b" in back and the "h" in house; it's not equally likely that you intended to say "I'm going tack inside the mouse." The AI system may be able to use what it knows of human language — and of your own voice or the contest in which you're speaking — to fill in the gaps intelligently.

But that's left to future research. For now you can read the team's work so far in the paper "Personalizing ASR for Dysarthric and Accented Speech with Limited Data," due to be presented at the Interspeech conference in Austria next month.

This is our best look yet at a Pixel 4 in the wild.
Google a few weeks ago showed a teaser image of the Pixel 4's rear, confirming the phone will feature a dual-lens camera placed inside into a rectangular camera module. Then, Google announced that the Pixel 4 would sport 3D face recognition support that's going to work differently than Face ID. Also, the company confirmed that the Project Soli radar system will be integrated into the top bezels. In the process, Google confirmed the Pixel 4 will not have a notch, as all the components needed for face recognition and radar-based gestures will occupy a lot of space on top of the phone.

—   Share news in the SOC. Networks

Topical videos:

usr: 4
This is interesting!