AI Glossary
The complete dictionary of Artificial Intelligence
On-Device Learning
Ability of wearable devices to perform continuous learning directly on the device without requiring constant cloud connection, adapting models to user's specificities.
PPG (Photoplethysmography)
Optical technique for measuring blood volume variations in body tissues, used in smartwatches for cardiac monitoring and SpO2 estimation.
ECG Monitoring
Real-time electrocardiographic monitoring through embedded sensors, detecting cardiac arrhythmias and other heart rhythm abnormalities directly on the wearable.
Activity Recognition
Automatic classification of user's physical activities (walking, running, swimming) in real-time from inertial and physiological sensor data.
Biometric Sensors
Set of sensors measuring unique user physiological parameters (heart rate, temperature, SpO2) for health monitoring and biometric authentication.
Neural Network Quantization
Process of reducing neural network weights and activations precision (typically from 32-bit to 8-bit or less) to optimize memory footprint and power consumption on wearables.
Heart Rate Variability (HRV)
Measurement of time variations between successive heartbeats, used as an indicator of stress, recovery, and general health status through Edge AI algorithms.
Sleep Pattern Analysis
Automatic detection of sleep phases (light, deep, REM) and rest quality analysis from accelerometer and physiological data collected during the night.
Low-Power Inference
Optimization of AI model execution to minimize energy consumption, crucial for the battery life of wearables running on battery for several days.
Wearable AI Chips
Specialized processors (NPU, TPU) specifically designed for accelerating AI workloads on portable devices with strict energy constraints.
Health Anomaly Detection
Automatic identification of abnormal patterns in physiological data (arrhythmias, hypoxia, fever) by AI models executed locally on the wearable.
Motion Sensors
Inertial sensors (3-axis accelerometer, 3-axis gyroscope, magnetometer) providing fundamental movement data for activity recognition and posture tracking.
Real-Time Health Monitoring
Continuous and instantaneous monitoring of vital parameters with local predictive analysis, enabling immediate alerts in case of detection of critical anomalies.
SpO2 Monitoring
Continuous measurement of blood oxygen saturation by multi-wavelength photoplethysmography, using AI algorithms to compensate for motion artifacts.