An abstract parabolic system-based physics-informed long short-term memory network for estimating breath alcohol concentration from transdermal alcohol biosensor data

Abstract

The problem of estimating breath alcohol concentration based on transdermal alcohol biosensor data is considered. Transdermal alcohol concentration provides a promising alternative to classical methods such as breathalyzers or drinking diaries. A physics-informed long short-term memory (LSTM) network with covariates for the solution of the estimation problem is developed. The data-driven nature of an LSTM is augmented with a first-principles physics-based population model for the diffusion of ethanol through the epidermal layer of the skin. The population model in an abstract parabolic framework appears as part of a regularization term in the loss function of the LSTM. While learning, the model is encouraged to both fit the data and to produce physically meaningful outputs. To deal with the high variation observed in the data, a mechanism for the uncertainty quantification of the estimates based on a recently discovered relation between Monte-Carlo dropout and Bayesian learning is used. The physics-based population model and the LSTM are trained and tested using controlled laboratory collected breath and transdermal alcohol data collected in four sessions from 40 orally dosed participants (50% female, ages 21–33 years, 35% BMI above 25.0) resulting in 256 usable drinking episodes partitioned into training and testing sets. Body measurement (e.g., BMI, hip to waist ratio, etc.), personal (e.g., sex, age, race, etc.), drinking behavior (e.g., frequent, rarely, etc.), and environmental (e.g., temperature, humidity, etc.) covariates were also collected from participants. The importance of various covariates in the estimation is investigated using Shapley values. It is shown that the physics-informed LSTM network can be successfully applied to drinking episodes from both the training and test sets and that the physics-based information leads to better generalization ability on new drinking episodes with the uncertainty quantification yielding credible bands that effectively capture the true signal. Compared to two machine learning models from previous studies, the proposed model reduces relative $L_2$ error in estimated breath alcohol concentration by 58 and 72% and relative peak error by 33 and 76%.

Appendix A List of Covariates

Name	Description
Session Type	Single, dual, or steady
Gender	Male or female
Ethnicity	African/African-American, Asian, caucasian, hispanics, or mixed
BMI	Body mass index
RM	Resting metabolism
SM	Skeletal muscle
Wais_Hip	waist-to-Hip ratio
AGE	Age of participants in years
ATE	Number of hours from eating to start of drinking session
WATER	Number of hours from time last drank water to time start drinking
CAL_LUN	Amount of calories for lunch
BPBase_S	Systolic blood pressure at rest
BPBase_D	Diastolic blood pressure at rest
BPPeak_S	Systolic blood pressure at peak
BPPeak_D	Diastolic blood pressure at peak
BPLow_S_S	Lowest systolic blood pressure reading before lunch
BPLow_S_D	Diastolic blood pressure reading at lowest systolic reading before lunch
BPLow_D_S	Systolic blood pressure reading at lowest diastolic reading before lunch
BPLow_D_D	Lowest diastolic blood pressure reading before lunch
28d_M	Past 28 days: maximum number of drinks on single day
28d_B	Number of binge days (5 drinks for women, 4 for men)
28d_T	Total number of drinks
90d_M	Past 90 days: maximum number of drinks on single day
90d_B	Number of binge days (5 drinks for women, 4 for men)
90d_T	Total number of drinks