arrow
Return

Explainable Artificial Intelligence Using Expressive Boolean Formulas

delete2023-11-24
delete6
delete
OA
AI
J
J. Kyle Brubaker
M
Martin J. A. Schuetz
G
Grant Salton
Z
Zhihuai Zhu
E
Elton Yechao Zhu
S
Serdar Kadıoğlu
H
Helmut G. Katzgraber
DOI:10.3390/make5040086delete
deleteOriginal
deleteShare
deleteSave
Abstract

Abstract

En 中文
We propose and implement an interpretable machine learning classification model for Explainable AI (XAI) based on expressive Boolean formulas. Potential applications include credit scoring and diagnosis of medical conditions. The Boolean formula defines a rule with tunable complexity (or interpretability) according to which input data are classified. Such a formula can include any operator that can be applied to one or more Boolean variables, thus providing higher expressivity compared to more rigid rule- and tree-based approaches. The classifier is trained using native local optimization techniques, efficiently searching the space of feasible formulas. Shallow rules can be determined by fast Integer Linear Programming (ILP) or Quadratic Unconstrained Binary Optimization (QUBO) solvers, potentially powered by special-purpose hardware or quantum devices. We combine the expressivity and efficiency of the native local optimizer with the fast operation of these devices by executing non-local moves that optimize over the subtrees of the full Boolean formula. We provide extensive numerical benchmarking results featuring several baselines on well-known public datasets. Based on the results, we find that the native local rule classifier is generally competitive with the other classifiers. The addition of non-local moves achieves similar results with fewer iterations. Therefore, using specialized or quantum hardware could lead to a significant speedup through the rapid proposal of non-local moves.
Keywords:
explainable AI
interpretable ML
Boolean formulas
stochastic local search
large neighborhood search
quantum computing
ILP
QUBO

Journal

M
Machine Learning and Knowledge Extraction
IF:
6
Papers:
900
Citations:
2.7K

Organization

C
California Institute of Technology
Scholars:
3.4K
Papers: 1.0K
Citations: 2.2W
Cited Papers

Cited Papers

Reductions in neural activity underlie behavioral components of repetition priming
err2005-07-31
err0
PREAI
errGagan S Wig; Scott T Grafton; Kathryn E Demos; William M Kelley
errShare
errSave
The Variational Quantum Eigensolver: A review of methods and best practices
err2022-11-01
err343
errOAAI
errTilly, Jules; Chen, Hongxiang; Cao, Shuxiang; Picozzi, Dario; Setia, Kanav; Li, Ying; Grant, Edward; Wossnig, Leonard; Rungger, Ivan; Booth, George H.; Tennyson, Jonathan
errShare
errSave
TRANSFORMATION OF THE INSECTICIDE TEFLUBENZURON BY MICROORGANISMS
err2007-02-06
err0
PREAI
errZoja I. Finkelstein; Boris P. Baskunov; Ivonne M. C. M. Rietjens; Marelle G. Boersma; Jacques Vervoort; Ludmila A. Golovleva
errShare
errSave
A data-driven approach to predict the success of bank telemarketing
err2014-06-01
err525
errOAAI
errMoro, Sergio; Cortez, Paulo; Rita, Paulo
errShare
errSave
Treatment of Parkinson's Disease with Bromocriptine
err1976-12-16
err0
PREAI
errAbraham Lieberman; Mark Kupersmith; Eli Estey; Menek Goldstein
errShare
errSave
Efficacy and safety of danoprevir plus sofosbuvir in GT 1, 2, 3, or 6 chronic hepatitis C patients with or without cirrhosis in China
err2021-06-18
err0
errOAAI
errShufang Pan; Kai Feng; Ping Huang; Yingfu Zeng; Liu Ke; Xiaodong Yang; Jing Liu; Chaoshuang Lin
errShare
errSave
Computational overhead of locality reduction in binary optimization problems
err2021-12-01
err5
errOAAI
errValiante, Elisabetta; Hernandez, Maritza; Barzegar, Amin; Katzgraber, Helmut G.
errShare
errSave
researcher View more