Abstract:The intersection of physics and machine learning has given rise to the physics-enhanced machine learning (PEML) paradigm, aiming to improve the capabilities and reduce the individual shortcomings of data- or physics-only methods. In this paper, the spectrum of physics-enhanced machine learning methods, expressed across the defining axes of physics and data, is discussed by engaging in a comprehensive exploration of its characteristics, usage, and motivations. In doing so, we present a survey of recent applications and developments of PEML techniques, revealing the potency of PEML in addressing complex challenges. We further demonstrate application of select such schemes on the simple working example of a single degree-of-freedom Duffing oscillator, which allows to highlight the individual characteristics and motivations of different `genres' of PEML approaches. To promote collaboration and transparency, and to provide practical examples for the reader, the code generating these working examples is provided alongside this paper. As a foundational contribution, this paper underscores the significance of PEML in pushing the boundaries of scientific and engineering research, underpinned by the synergy of physical insights and machine learning capabilities.

What problem does this paper attempt to address?

### Problems Addressed by the Paper This paper explores the spectrum of Physics-Enhanced Machine Learning (PEML) in structural mechanics applications, aiming to address the following issues: 1. **Limitations of Data-Driven Models**: - Data-driven models are often limited to the specific domain of data collection, lacking generalization capability. - Collecting data across the full range of environmental and operational conditions is often impractical, especially in large-scale engineering systems. - Limited data samples may lead to models failing to accurately capture intrinsic relationships, resulting in misleading extrapolation outcomes. 2. **Limitations of Physical Models**: - Pure physical models may be overly simplified, failing to accurately describe the nonlinear and non-stationary behavior of complex systems. - Physical models have limitations in handling uncertainties and modeling errors. 3. **Combining the Advantages of Physics and Data**: - The PEML approach aims to overcome the shortcomings of purely data-driven or purely physical methods by embedding physical knowledge into machine learning models. - This combination can enhance the interpretability and generalization capability of models, making them more suitable for scientific research and engineering applications. 4. **Classification and Understanding of PEML Methods**: - The paper classifies PEML methods through a two-dimensional spectrum, with one axis representing the degree of reliance on physical models and the other axis representing the degree of reliance on data. - Through this classification, researchers and users can choose appropriate PEML methods based on the characteristics and needs of specific problems. 5. **Demonstrating the Application of PEML Methods**: - The paper demonstrates the characteristics and potential of different types of PEML methods through an example of a single-degree-of-freedom Duffing oscillator. - It provides the code for generating these examples to promote transparency and collaboration. ### Main Contributions - **Theoretical Framework**: A classification framework for Physics-Enhanced Machine Learning methods is proposed, helping researchers better understand and choose appropriate methods. - **Application Example**: The application of PEML methods in structural mechanics is demonstrated through specific examples (e.g., Duffing oscillator). - **Code Sharing**: The code for generating examples is provided to promote collaboration and transparency within the research community. In summary, this paper aims to push the boundaries of scientific and engineering research by integrating physical knowledge and machine learning techniques, addressing the limitations of traditional methods in handling complex systems.

Discussing the Spectrum of Physics-Enhanced Machine Learning; a Survey on Structural Mechanics Applications

Physics-Enhanced Machine Learning: a position paper for dynamical systems investigations

When Physics Meets Machine Learning: A Survey of Physics-Informed Machine Learning

Physics-Informed Machine Learning: A Survey on Problems, Methods and Applications

Machine Learning Methodology Review for Computational Electromagnetics

Physics-Informed Machine Learning for Modeling and Control of Dynamical Systems

Physics-informed machine learning and its structural integrity applications: state of the art

Machine Learning and Physics: A Survey of Integrated Models

The Application of Physics-Informed Machine Learning in Multiphysics Modeling in Chemical Engineering

Physics-informed machine learning

Physics-Informed Extreme Learning Machine framework for solving linear elasticity mechanics problems

Machine Learning with Physics Knowledge for Prediction: A Survey

A review of physics-based machine learning in civil engineering

Perspective: Machine learning in experimental solid mechanics

Physics-Informed Machine Learning for Seismic Response Prediction OF Nonlinear Steel Moment Resisting Frame Structures

Integrating Machine Learning with Physics-Based Modeling

Revolutionizing physics: a comprehensive survey of machine learning applications

Recent Advances and Applications of Machine Learning in Experimental Solid Mechanics: A Review

Physics-informed machine learning for Structural Health Monitoring

Physics-Informed Machine Learning—An Emerging Trend in Tribology