In this article we analyze two datasets of physiological signals and psychological metadata for cognitive load inference. Multiple approaches for cognitive load inference with machine learning techniques are presented. The results of this study motivate the use of approximate mobile computing for cognitive load inference using mobile and wearable devices.
COBISS.SI-ID: 17709571
In this article we present RICERCANDO, a tool for mining mobile broadband data that is based on a specially constructed machine learning pipeline. The tool can be used to detect anomalies and identify causes for certain network behaviours. The machine learning pipeline described in the article opens opportunities for the use of approximate computing for on-device network measurement data processing.
COBISS.SI-ID: 17175043
In this article we examine the use of sensor data collected from IoT devices in an office-like environment for continuous authentication. By showing that machine learning models built on such data can indeed provide promising authentication performance (99.3% accuracy in our experiments), we open space for the use of approximate mobile computing for distributed learning for the purpose of building authentication models.
COBISS.SI-ID: 44888067