Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Interpretable Hybrid Machine Learning Models Using FOLD-R++ and Answer Set Programming

Sanne Wielinga, Jesse Heyninck

TL;DR

This work tackles the interpretability gap in high-stakes medical prediction by marrying symbolic reasoning with neural and non-neural classifiers. It introduces a model-agnostic hybrid that uses FOLD-R++-generated ASP rules to selectively correct low-confidence ML predictions and generate human-readable explanations. Empirical results on five medical datasets show statistically significant gains in accuracy and F1 for several classifiers, most notably SVM, demonstrating that symbolic rules can enhance performance without altering ML internals. The approach offers concrete interpretability through proof-tree explanations and has practical implications for transparent clinical decision support, while acknowledging computational overhead and variability in rule induction as areas for future improvement.

Abstract

Machine learning (ML) techniques play a pivotal role in high-stakes domains such as healthcare, where accurate predictions can greatly enhance decision-making. However, most high-performing methods such as neural networks and ensemble methods are often opaque, limiting trust and broader adoption. In parallel, symbolic methods like Answer Set Programming (ASP) offer the possibility of interpretable logical rules but do not always match the predictive power of ML models. This paper proposes a hybrid approach that integrates ASP-derived rules from the FOLD-R++ algorithm with black-box ML classifiers to selectively correct uncertain predictions and provide human-readable explanations. Experiments on five medical reveal statistically significant performance gains in accuracy and F1 score. This study underscores the potential of combining symbolic reasoning with conventional ML to achieve high interpretability without sacrificing accuracy

Interpretable Hybrid Machine Learning Models Using FOLD-R++ and Answer Set Programming

TL;DR

This work tackles the interpretability gap in high-stakes medical prediction by marrying symbolic reasoning with neural and non-neural classifiers. It introduces a model-agnostic hybrid that uses FOLD-R++-generated ASP rules to selectively correct low-confidence ML predictions and generate human-readable explanations. Empirical results on five medical datasets show statistically significant gains in accuracy and F1 for several classifiers, most notably SVM, demonstrating that symbolic rules can enhance performance without altering ML internals. The approach offers concrete interpretability through proof-tree explanations and has practical implications for transparent clinical decision support, while acknowledging computational overhead and variability in rule induction as areas for future improvement.

Abstract

Machine learning (ML) techniques play a pivotal role in high-stakes domains such as healthcare, where accurate predictions can greatly enhance decision-making. However, most high-performing methods such as neural networks and ensemble methods are often opaque, limiting trust and broader adoption. In parallel, symbolic methods like Answer Set Programming (ASP) offer the possibility of interpretable logical rules but do not always match the predictive power of ML models. This paper proposes a hybrid approach that integrates ASP-derived rules from the FOLD-R++ algorithm with black-box ML classifiers to selectively correct uncertain predictions and provide human-readable explanations. Experiments on five medical reveal statistically significant performance gains in accuracy and F1 score. This study underscores the potential of combining symbolic reasoning with conventional ML to achieve high interpretability without sacrificing accuracy

Paper Structure

This paper contains 24 sections, 1 figure, 1 table.

Table of Contents

  1. Introduction
  2. Related Work
  3. Interpretability and Related Approaches
  4. Hybrid Models
  5. Background
  6. Symbolic Logic and Answer Set Programming
  7. FOLD-R++ Algorithm
  8. Traditional ML Methods
  9. Methods
  10. Data Preparation
  11. FOLD-R++
  12. ML Models
  13. Hybrid Model Implementation
  14. Experimental Setup
  15. Results
  16. ...and 9 more sections

Figures (1)

  • Figure 1: Conceptual diagram of the hybrid model architecture. Input data is processed by both the black-box ML model and the ASP rule-based component. The confidence score determines whether the ML prediction is accepted directly or corrected by the ASP rules.