Amazon Technologies, Inc.

Michael Thomas Peterson

Carlo Murgia

Aditya Sharadchandra Joshi

Ilya Sokolov

Gengshen Fu

Aaron Challenner

Rong Chen

Cody Jacques

Stefan M Bradstreet

Andrew Morris Werchniak

Aansh Shah

Raphael Petegrosso

A system performs federated learning and retraining of a machine learning model used for processing audio detected by a user device. The system uses both gradient data (which may correspond to false-rejects) and audio data (which may correspond to false-positives) received from devices. The system may also use a teacher model to produce labels for data in an automated fashion, thus allowing retraining to happen in an unsupervised manner.

Federated learning for audio processing

A device is configured to detect multiple different wakewords. A device may operate a joint encoder that operates on audio data to determine encoded audio data. The device may operate multiple different decoders which process the encoded audio data to determine if a wakeword is detected. Each decoder may correspond to a different wakeword. The decoders may use fewer computing resources than the joint encoder, allowing for the device to more easily perform multiple wakeword processing. Enabling/disabling wakeword(s) may involve the reconfiguring of a wakeword detector to add/remove data for respective decoder(s). Specific decoders may be activated/deactivated depending on device context, thereby efficiently managing device resources.

Audio detection

Techniques for encoding audio data with metadata are described. In an example, a device receives audio data corresponding to audio detected by a microphone and receives metadata associated with the audio. The device generates encoded data based at least in part on encoding the audio data with the metadata. The encoding involves replacing a portion of the audio data with the metadata, such that the encoded data includes the metadata and a remaining portion of the audio data. The device sends the encoded data to an audio processing application.

Encoding audio metadata in an audio frame

Growing community of inventors

Watertown, MA, United States of America

Michael Thomas Peterson