DATA COMPRESSION FOR ELECTRONIC PERSONAL HEALTH DEVICES USING AN ENCODER-DECODER ARTIFICIAL INTELLIGENCE MODEL

The invention provides a symmetric encoder-decoder AI model for data compression on low-resource personal care devices and devises technologies for user task-specific model extensions as well as on-device deployment and operation. A method for compressing data on an electronic personal health device, such as an electronic toothbrush or an electronic shaver, is provided, comprising capturing sensor data using a sensor, particularly an inertial measurement unit, of the personal health device; compressing the sensor data on the personal health device using an encoder model, particularly by generating one or more latent vectors in a latent space based at least in part on the captured sensor data, and storing the compressed sensor data on a non-volatile storage medium of the personal health device. The compressed data may be reconstructed using a decoder on a smartphone or a cloud computing environment.