Enhanced Detection of Transdermal Alcohol Levels Using Hyperdimensional Computing on Embedded Devices

TL;DR

This work investigates on-device detection of transdermal alcohol levels using hyperdimensional computing (HDC) to enable practical just-in-time interventions. By evaluating multiple HDC encodings and learning models, the authors demonstrate that a generic encoding with a density-based ensemble and a RefineHD classifier achieves about 89 percent accuracy on chronologically ordered data, outperforming prior ML approaches. The approach runs efficiently on a Raspberry Pi, with classification times around 0.34 seconds per 10-second window, supporting real-time, on-device JITAIs for alcohol monitoring. The study also shows that temporal structuring of data improves accuracy by a notable margin, underscoring the value of time-series-aware HDC in resource-constrained environments. The results suggest a viable path toward low-power, privacy-preserving, on-device alcohol monitoring using consumer wearables and smartphones.

Abstract

Alcohol consumption has a significant impact on individuals' health, with even more pronounced consequences when consumption becomes excessive. One approach to promoting healthier drinking habits is implementing just-in-time interventions, where timely notifications indicating intoxication are sent during heavy drinking episodes. However, the complexity or invasiveness of an intervention mechanism may deter an individual from using them in practice. Previous research tackled this challenge using collected motion data and conventional Machine Learning (ML) algorithms to classify heavy drinking episodes, but with impractical accuracy and computational efficiency for mobile devices. Consequently, we have elected to use Hyperdimensional Computing (HDC) to design a just-in-time intervention approach that is practical for smartphones, smart wearables, and IoT deployment. HDC is a framework that has proven results in processing real-time sensor data efficiently. This approach offers several advantages, including low latency, minimal power consumption, and high parallelism. We explore various HDC encoding designs and combine them with various HDC learning models to create an optimal and feasible approach for mobile devices. Our findings indicate an accuracy rate of 89\%, which represents a substantial 12\% improvement over the current state-of-the-art.

Figures (3)

• Figure 1: Hyperdimensional classification workflow.
• Figure 2: Performance evaluation of the generic encoding with different n-gram sizes, with randomly shuffled and chronologically ordered data.
• Figure 3: ROC curve evaluation of the ensemble Generic encoding and RefineHD model with different learning rates, with randomly shuffled data.