A new Google AI tool filters noisy wearable sensor data to identify key candidate biomarkers. The system reduces manual data labeling by automating the selection of physiological signals. This streamlines how researchers isolate health indicators from continuous streams. Practitioners can now accelerate the discovery phase for digital health diagnostics without exhaustive manual review.