Diverse Explanations From Data-Driven and Domain-Driven Perspectives in the Physical Sciences
Abstract
Machine learning methods have been remarkably successful in material science, providing novel scientific insights, guiding future laboratory experiments, and accelerating materials discovery. Despite the promising performance of these models, understanding the decisions they make is also essential to ensure the scientific value of their outcomes. However, there is a recent and ongoing debate about the diversity of explanations, which potentially leads to scientific inconsistency. This Perspective explores the sources and implications of these diverse explanations in ML applications for physical sciences. Through three case studies in materials science and molecular property prediction, we examine how different models, explanation methods, levels of feature attribution, and stakeholder needs can result in varying interpretations of ML outputs. Our analysis underscores the importance of considering multiple perspectives when interpreting ML models in scientific contexts and highlights the critical need for scientists to maintain control over the interpretation process, balancing data-driven insights with domain expertise to meet specific scientific needs. By fostering a comprehensive understanding of these inconsistencies, we aim to contribute to the responsible integration of eXplainable Artificial Intelligence (XAI) into physical sciences and improve the trustworthiness of ML applications in scientific discovery.
AI-Generated Overview
-
Research Focus: The study investigates the diverse explanations arising from data-driven and domain-driven machine learning (ML) models in the physical sciences, focusing on materials science applications.
-
Methodology: The authors conduct three case studies involving materials science and molecular property prediction, analyzing how various models, explanation methods, and stakeholder perspectives contribute to inconsistencies in the interpretation of ML outputs.
-
Results: Different ML models and explanation methods produce varying interpretations of the same data, demonstrating that both data-driven and domain-driven explanations can conflict, leading to inconsistencies and challenges in model trustworthiness.
-
Key Contribution(s): The paper highlights the importance of a comprehensive understanding of diverse explanations in ML applications to foster responsible integration of eXplainable Artificial Intelligence (XAI) in the physical sciences, advocating for a scientist-centric approach to interpretation.
-
Significance: By addressing the discrepancies in interpretations and the implications for scientific discovery, the study aims to enhance the reliability of ML applications and deepen the understanding of complex physical phenomena, ultimately benefiting material research.
-
Broader Applications: The findings extend beyond materials science, suggesting pathways for improving the interpretability and trustworthiness of ML models in various scientific disciplines, including nanotechnology and drug discovery, where robust model interpretation is critical for informed decision-making.