Artificial intelligence lets your smart assistant know the direction of your voice


Artificial intelligence may soon be a companion to the smart assistant in an important feature, as Carnegie Mellon University researchers have developed a machine learning model that estimates the direction in which the sound is coming from, and indicates your intention to wake the device or not without the need for a special phrase or gesture, and the method depends on the characteristics inherent in the sound. While hesitating around the room.

According to "engadget", the system realizes that the first, loudest and clearest sound is always the voice directed directly at a specific subject, and anything else tends to be quieter and delayed. The model also realizes that human speech frequencies vary depending on the direction you are facing, where Low frequencies tend to be multi-directional.

The researchers added that this method is simple, software-dependent and does not require audio data to be sent to the cloud, and it may take some time before you see the technology in use, although the team has released code and data publicly to help others build their work.

It's also easy to see where this might lead, so you can tell a smart assistant to play music without using a nudge or setting a bunch of other connected devices.

It may also aid privacy by requiring your physical presence while avoiding the need to have cameras to detect your gaze

Previous Post Next Post