Related Experiment Video
Updated: Jul 29, 2025
08:20
Author Spotlight: AI-Driven Trypanosome Species Detection from Microscopic Images
Published on: October 27, 2023
1.5K
Predicting facility-based delivery in Zanzibar: The vulnerability of machine learning algorithms to adversarial
Yi-Ting Tsai1, Isabel R Fulcher2,3, Tracey Li4
1Department of Biostatistics, Harvard Chan School of Public Health, Boston, USA.
Heliyon
|May 26, 2023
Summary
Machine learning models identifying at-risk mothers for home delivery are vulnerable to adversarial attacks. Data manipulation, especially on previous delivery location, can alter predictions, impacting program effectiveness in sub-Saharan Africa.
Area of Science:
- Machine Learning in Public Health
- Maternal Health Interventions
- Data Security in Healthcare
Background:
- Community health worker (CHW)-led programs improve maternal health in sub-Saharan Africa.
- Mobile device adoption enables real-time machine learning for identifying high-risk pregnancies.
- Adversarial attacks, or falsified data, pose a threat to the accuracy of these predictive models.
Purpose of the Study:
- To evaluate the vulnerability of a machine learning algorithm to adversarial attacks.
- To assess the impact of data manipulation on the prediction of home-based delivery risk.
Main Methods:
- Utilized data from the Uzazi Salama program in Zanzibar (2016-2019).
- Developed a prediction model using LASSO regularized logistic regression.
- Applied "One-At-a-Time (OAT)" adversarial attacks to various input variables (binary, categorical, ordinal, continuous).
Main Results:
- Adversarial attacks significantly altered prediction outcomes.
- The "previous delivery location" variable showed the highest vulnerability.
- Predictions changed by 55.65% and 37.63% when manipulating this variable between facility and home delivery.
Conclusions:
- The study highlights the susceptibility of maternal health predictive algorithms to adversarial manipulation.
- Understanding these vulnerabilities is crucial for developing robust data monitoring strategies.
- Ensuring data integrity is essential for effective targeting of at-risk women by community health workers.
Keywords:
Adversarial attackCommunity health worker interventionDigital healthFacility deliveryMachine learningMaternal healthMore Related Videos
Related Concept Videos
01:19Survival Tree
125
Survival trees are a non-parametric method used in survival analysis to model the relationship between a set of covariates and the time until an event of interest occurs, often referred to as the "time-to-event" or "survival time." This method is particularly useful when dealing with censored data, where the event has not occurred for some individuals by the end of the study period, or when the exact time of the event is unknown.
Building a Survival Tree
Constructing a...
Building a Survival Tree
Constructing a...
125
01:18Steps in Outbreak Investigation
155
In the ever-evolving field of public health, statistical analysis serves as a cornerstone for understanding and managing disease outbreaks. By leveraging various statistical tools, health professionals can predict potential outbreaks, analyze ongoing situations, and devise effective responses to mitigate impact. For that to happen, there are a few possible stages of the analysis:
155
01:25Distribution Reliability and Automation
136
Distribution reliability in electrical power systems is critical for ensuring an uninterrupted power supply to consumers at minimal cost. According to IEEE Standard Terms, reliability is the probability that a device will function without failure over a specified time period or amount of usage. For electric power distribution, this translates to maintaining continuous power supply and addressing customer concerns over power outages. Several indices, as defined by IEEE Standard 1366-2012, are...
136
01:03Prediction Intervals
2.3K
The interval estimate of any variable is known as the prediction interval. It helps decide if a point estimate is dependable.
However, the point estimate is most likely not the exact value of the population parameter, but close to it. After calculating point estimates, we construct interval estimates, called confidence intervals or prediction intervals. This prediction interval comprises a range of values unlike the point estimate and is a better predictor of the observed sample value, y.
However, the point estimate is most likely not the exact value of the population parameter, but close to it. After calculating point estimates, we construct interval estimates, called confidence intervals or prediction intervals. This prediction interval comprises a range of values unlike the point estimate and is a better predictor of the observed sample value, y.
2.3K