Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Designing User-Centric Behavioral Interventions to Prevent Dysglycemia with Novel Counterfactual Explanations

Asiful Arefeen, Hassan Ghasemzadeh

TL;DR

ExAct is unique in integrating prior knowledge about user preferences of feasible explanations into the process of counterfactual generation, and surpasses the state-of-the-art techniques for generating counterfactual explanations for chronic disease prevention and management.

Abstract

Monitoring unexpected health events and taking actionable measures to avert them beforehand is central to maintaining health and preventing disease. Therefore, a tool capable of predicting adverse health events and offering users actionable feedback about how to make changes in their diet, exercise, and medication to prevent abnormal health events could have significant societal impacts. Counterfactual explanations can provide insights into why a model made a particular prediction by generating hypothetical instances that are similar to the original input but lead to a different prediction outcome. Therefore, counterfactuals can be viewed as a means to design AI-driven health interventions to not only predict but also prevent adverse health outcomes such as blood glucose spikes, diabetes, and heart disease. In this paper, we design \textit{\textbf{ExAct}}, a novel model-agnostic framework for generating counterfactual explanations for chronic disease prevention and management. Leveraging insights from adversarial learning, ExAct characterizes the decision boundary for high-dimensional data and performs a grid search to generate actionable interventions. ExAct is unique in integrating prior knowledge about user preferences of feasible explanations into the process of counterfactual generation. ExAct is evaluated extensively using four real-world datasets and external simulators. With $82.8\%$ average validity in the simulation-aided validation, ExAct surpasses the state-of-the-art techniques for generating counterfactual explanations by at least $10\%$. Besides, counterfactuals from ExAct exhibit at least $6.6\%$ improved proximity compared to previous research.

Designing User-Centric Behavioral Interventions to Prevent Dysglycemia with Novel Counterfactual Explanations

TL;DR

ExAct is unique in integrating prior knowledge about user preferences of feasible explanations into the process of counterfactual generation, and surpasses the state-of-the-art techniques for generating counterfactual explanations for chronic disease prevention and management.

Abstract

Monitoring unexpected health events and taking actionable measures to avert them beforehand is central to maintaining health and preventing disease. Therefore, a tool capable of predicting adverse health events and offering users actionable feedback about how to make changes in their diet, exercise, and medication to prevent abnormal health events could have significant societal impacts. Counterfactual explanations can provide insights into why a model made a particular prediction by generating hypothetical instances that are similar to the original input but lead to a different prediction outcome. Therefore, counterfactuals can be viewed as a means to design AI-driven health interventions to not only predict but also prevent adverse health outcomes such as blood glucose spikes, diabetes, and heart disease. In this paper, we design \textit{\textbf{ExAct}}, a novel model-agnostic framework for generating counterfactual explanations for chronic disease prevention and management. Leveraging insights from adversarial learning, ExAct characterizes the decision boundary for high-dimensional data and performs a grid search to generate actionable interventions. ExAct is unique in integrating prior knowledge about user preferences of feasible explanations into the process of counterfactual generation. ExAct is evaluated extensively using four real-world datasets and external simulators. With average validity in the simulation-aided validation, ExAct surpasses the state-of-the-art techniques for generating counterfactual explanations by at least . Besides, counterfactuals from ExAct exhibit at least improved proximity compared to previous research.
Paper Structure (35 sections, 1 theorem, 14 equations, 7 figures, 12 tables, 2 algorithms)

This paper contains 35 sections, 1 theorem, 14 equations, 7 figures, 12 tables, 2 algorithms.

Table of Contents

  1. Introduction
  2. Problem statement
  3. Problem solution
  4. Experimental Setup
  5. Results
  6. CONCLUSION
  7. Resource Specifications
  8. Dataset Descriptions
  9. Heart Disease
  10. PimaDM
  11. Nutrition Absorption Dataset
  12. OhioT1DM
  13. Model Specifications
  14. Heart Disease
  15. Classifier
  16. ...and 20 more sections

Key Result

Lemma 1

For a point $x$ and a non-empty set $S$ in a metric space, the distance from $x$ to $S$, denoted as $d(x, S)$, is equal to the infimum of the distances between $x$ and all points in $S$ i.e. $d(x, S) = \inf\{d(x, y) : y \in S\}$

Figures (7)

  • Figure 1: Proposed approach to solve the optimization problem. (a) A classifier (black dashed line) is trained to identify normal (green) vs. abnormal (red) classes. (b) Many adversarial critical samples (borderline instances in blue and yellow) are generated along the decision boundary to approximate the decision hyperplane. (c) The real samples can be removed once substantial number of critical samples are in place. For a test sample outside of the normal region, predicted class can be toggled following an infinite number of trajectories. However, ExAct follows the path that requires minimal changes or emphasizes users preferences and restricts some features while flipping the class.
  • Figure 2: The autoencoders produce borderline instances (blue and yellow), while the addition of a bisection algorithm adds finesse to them (teal triangles). The numbers are class probabilities.
  • Figure 3: Constrained Intervention involves altering the predicted class of the test sample with as few one-dimensional moves as possible. Considering the blue volume as the normal region and its surface as the decision hyperplane, (a and b) the test sample has two of the three feature values within the range of the decision hyperplane. (c) Hence, a move along y-axis is sufficient to place the sample within the normal zone.
  • Figure 4: Outputs of the autoencoders and the bisection algorithm, i.e., the borderline instances, are shown across three plots. The plots display six out of the seven features, grouped as (a) pre-meal BGL (i.e. SBGL), insulin intake, CHO, (b) CHO, fiber intake, CHO composition, and (c) DFB (i.e. time elapsed since last insulin), fiber intake, insulin intake. This visualization demonstrates the precise placement of critical instances between the quasi-hyperglycemic and normoglycemic samples.
  • Figure 5: Proximity-invalidity trade-off across all methods.
  • ...and 2 more figures

Theorems & Definitions (2)

  • Lemma 1
  • proof