Today Apple published an entry in its Machine Learning Journal that explains how Siri works on the HomePod.
Optimizing Siri on HomePod in Far-Field Settings
The entry, written by the audio software engineering team with the Siri speech team, reveals why Apple has been focusing a lot on machine learning lately. A far-field setting means people using Siri on the HomePod from a distance, like the opposite end of the room. Specifically, the teams discuss two things:
- Mask-based multichannel filtering using deep learning to remove echo and background noise
- Unsupervised learning to separate simultaneous sound sources and trigger-phrase based stream selection to eliminate interfering speech
Although the journal entry is highly technical in nature, it’s a fascinating glimpse into the technology underpinning the HomePod.