Google Neural Network Can Isolate Individual Voices in Videos

The bleeding edge of computer science these days is all about making computers more like humans. We’re using neural networks to help machines recognize objects, play games, and even speak in a more realistic way. In a new feat of machine learning magic, Google Research has developed a system that can replicate the “cocktail party effect,” where your brain focuses on a single audio source in a crowded room. The results are impressive — almost worryingly so.

Google calls this technique “Looking to Listen” because it watches videos with multiple speakers to split up the audio—it uses both auditory and visual signals, just like your brain does. There’s nothing special about these videos, either. They’re just videos with a single audio track consisting of more than one person.

To build a tool capable of this, Google started with 100,000 samples of high-quality lectures and talks from YouTube. Engineers chopped up the videos to get segments of clean speech with clearly visible speakers and no background noise. That left Google Research with 2,000 hours of video consisting of a single person speaking (they call this the AVSpeech data set). The trick was using these clean samples to create “fake” cocktail parties. The researchers combined the videos, so multiple people were speaking. That’s the data Google used to train its neural network.

Like many other Google Research breakthroughs, this one used a convolutional neural network. The input to the network consists of visual features of the speakers as well as the spectrogram of the video’s soundtrack. By processing the video, the network learns how to separate the video into a “time frequency mask” for each speaker. The output mask is matched up with the audio input spectrogram to generate separate audio tracks.

With the training done, Google unleashed the network on new videos. As you can see in Google’s examples, this works surprisingly well. The Looking to Listen model can identify what audio is coming from a speaker, and filter out everything else. This technology could have applications in video conferencing, hearing aids, and video surveillance.

On that last point, this technology could be so powerful that it’s not hard to imagine scenarios where it’s abused. With future speed and accuracy improvements, an observer could pick out your voice on a crowded street to find out what you said. There’s no indication Google has any intention of doing that, but it’s not alone in doing neural network research.

Continue reading

Twitter’s Internal Research Confirms Its Algorithm Favors Right-Wing Voices

In recent years, conservative politicians and activists have railed against "cancel culture", claiming Twitter and other social media outlets are biased against them—hence, the existence of right-wing alternatives like Gab, Parler, and Trump's new Truth Social. Twitter has just published some internal research that suggests the opposite. According to the paper, its algorithm actually favors right-wing voices.

Amazon’s Alexa Will Soon be Able to Use a Dead Person’s Voice

Amazon unveiled the technology at its annual re:MARS event, which centers machine learning, automation, robotics, and space.

Red Moon Rising: NASA Admin Voices Fears of China Attempting Lunar Takeover

Nelson isn't the only US government official who's raised similar concerns that China means to attempt a military takeover of the Moon.

Google Begins Testing Wi-Fi Calls for Google Voice

After years of neglect, Google is again working to improve Google Voice by updating its apps and rolling out new features — and one long-awaited feature has just entered testing.