A Novel XAI-Enhanced Quantum Adversarial Networks for Velocity Dispersion Modeling in MaNGA Galaxies

2025-10-28

AI Summary

Addressing the challenge of simultaneously achieving high predictive accuracy, robustness, and interpretability in quantum machine learning, this paper proposes an eXplainable AI-enhanced Quantum Adversarial Network (XQAN) for modeling stellar velocity dispersion in MaNGA galaxies. Methodologically, XQAN integrates a quantum neural network (QNN) with a classical deep learning architecture into an end-to-end hybrid model; it introduces a LIME-driven interpretability constraint and synergistically couples quantum generative adversarial training with self-supervised learning to jointly optimize prediction performance and model transparency. Experimental results demonstrate that XQAN achieves RMSE = 0.27, MAE = 0.21, and R² = 0.59 on the regression task—substantially outperforming baseline methods—while exhibiting strong robustness and physically grounded interpretability. This work establishes a novel paradigm for quantum-classical hybrid modeling in astrophysics.

Technology Category

• Machine Learning: Quantum Machine Learning
• Humans and AI: Explainable AI (XAI) for Human Understanding
• Computer Vision: Generative Adversarial Networks (GANs) for Vision

Application Category

• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Current quantum machine learning approaches often face challenges balancing predictive accuracy, robustness, and interpretability. To address this, we propose a novel quantum adversarial framework that integrates a hybrid quantum neural network (QNN) with classical deep learning layers, guided by an evaluator model with LIME-based interpretability, and extended through quantum GAN and self-supervised variants. In the proposed model, an adversarial evaluator concurrently guides the QNN by computing feedback loss, thereby optimizing both prediction accuracy and model explainability. Empirical evaluations show that the Vanilla model achieves RMSE = 0.27, MSE = 0.071, MAE = 0.21, and R^2 = 0.59, delivering the most consistent performance across regression metrics compared to adversarial counterparts. These results demonstrate the potential of combining quantum-inspired methods with classical architectures to develop lightweight, high-performance, and interpretable predictive models, advancing the applicability of QML beyond current limitations.

Problem

• Enhancing predictive accuracy in quantum machine learning models
• Improving model robustness and interpretability through XAI integration
• Optimizing velocity dispersion modeling for MaNGA galaxies

Innovation

• Hybrid quantum neural network with classical layers
• Adversarial evaluator using LIME-based interpretability
• Quantum GAN and self-supervised variants extension