Tiredness Detection for a Robotic Mobility Aid

The proportion of elderly people in society is predicted to continue to rise in the coming decades. Mobility is a key aspect of many daily activities, but falls become an increasingly significant health risk with age. With the COVID-19 pandemic, many elderly users prefer or require assistive devices, rather than human support, in walking and carrying out daily tasks. However, prior work has shown that when using passive assistive mobility devices, fall risks can actually increase. This presents an opportunity for assistive robots to help maintain and improve the mobility of elderly users, with an additional emphasis on safety, made possible through sensing capabilities. In this paper, we present a computer vision system that detects the eye blink and face angle patterns for exhibiting signs of tiredness. In addition to the frame-based detection, we also introduce a time-window collation with a machine learning classifier. The system proposed here is critical in monitoring the user, performing real-time detection, and recommending they take a break if tiredness is detected. The overall system architecture and algorithmic details are presented, then a series of experiments are conducted to validate the performance of the approach. © 2022 IEEE.

Variational Disentanglement for Rare Event Modeling

Combining the increasing availability and abundance of healthcare data and the current advances in machine learning methods have created renewed opportunities to improve clinical decision support systems. However, in healthcare risk prediction applications, the proportion of cases with the condition (label) of interest is often very low relative to the available sample size. Though very prevalent in healthcare, such imbalanced classification settings are also common and challenging in many other scenarios. So motivated, we propose a variational disentanglement approach to semi-parametrically learn from rare events in heavily imbalanced classification problems. Specifically, we leverage the imposed extreme-distribution behavior on a latent space to extract information from low-prevalence events, and develop a robust prediction arm that joins the merits of the generalized additive model and isotonic neural nets. Results on synthetic studies and diverse real-world datasets, including mortality prediction on a COVID-19 cohort, demonstrate that the proposed approach outperforms existing alternatives.