Explainable Machine Learning for Scientific Insights and Discoveries

Abstract

Machine learning methods have been remarkably successful for a wide range of application areas in the extraction of essential information from data. An exciting and relatively recent development is the uptake of machine learning in the natural sciences, where the major goal is to obtain novel scientific insights and discoveries from observational or simulated data. A prerequisite for obtaining a scientific outcome is domain knowledge, which is needed to gain explainability, but also to enhance scientific consistency. In this article we review explainable machine learning in view of applications in the natural sciences and discuss three core elements which we identified as relevant in this context: transparency, interpretability, and explainability. With respect to these core elements, we provide a survey of recent scientific works that incorporate machine learning and the way that explainable machine learning is used in combination with domain knowledge from the application areas.

References in corpus (50)

• Discovering governing equations from data: Sparse identification of nonlinear dynamical systems
• Towards A Rigorous Science of Interpretable Machine Learning
• Methods for Interpreting and Understanding Deep Neural Networks
• Machine learning and the physical sciences
• Interpretable machine learning: definitions, methods, and applications
• European Union regulations on algorithmic decision-making and a "right to explanation"
• Theory-guided Data Science: A New Paradigm for Scientific Discovery from Data
• Kernel methods in machine learning
• Quantum-Chemical Insights from Deep Tensor Neural Networks
• Deep learning for universal linear embeddings of nonlinear dynamics
• Hidden Physics Models: Machine Learning of Nonlinear Partial Differential Equations
• Explainable Artificial Intelligence: Understanding, Visualizing and Interpreting Deep Learning Models
• Unmasking Clever Hans Predictors and Assessing What Machines Really Learn
• Physics Informed Deep Learning (Part I): Data-driven Solutions of Nonlinear Partial Differential Equations
• Informed Machine Learning -- A Taxonomy and Survey of Integrating Knowledge into Learning Systems
• Explaining Explanations in AI
• Learning to Segment Object Candidates
• Discovering Phase Transitions with Unsupervised Learning
• Machine Learning of Linear Differential Equations using Gaussian Processes
• Discovering physical concepts with neural networks
• Discovering Phases, Phase Transitions and Crossovers through Unsupervised Machine Learning: A critical examination
• Physics Informed Deep Learning (Part II): Data-driven Discovery of Nonlinear Partial Differential Equations
• Accelerating Eulerian Fluid Simulation With Convolutional Networks
• Syntax-Directed Variational Autoencoder for Structured Data
• Attention-Based Deep Neural Networks for Detection of Cancerous and Precancerous Esophagus Tissue on Histopathological Slides
• Sparse Dynamics for Partial Differential Equations
• A Compositional Object-Based Approach to Learning Physical Dynamics
• Learning physical descriptors for materials science by compressed sensing
• Causality for Machine Learning
• Physics-Aware Gaussian Processes in Remote Sensing
• Scientific Domain Knowledge Improves Exoplanet Transit Classification with Deep Learning
• Probing hidden spin order with interpretable machine learning
• Learning Physical Intuition of Block Towers by Example
• Prediction of Reynolds Stresses in High-Mach-Number Turbulent Boundary Layers using Physics-Informed Machine Learning
• Learning multiple order parameters with interpretable machines
• Attend and Predict: Understanding Gene Regulation by Selective Attention on Chromatin
• To Fall Or Not To Fall: A Visual Approach to Physical Stability Prediction
• Manifold learning for parameter reduction
• Conditioning by adaptive sampling for robust design
• Interpretable machine learning for inferring the phase boundaries in a nonequilibrium system
• Learning A Physical Long-term Predictor
• Design by adaptive sampling
• Automated, predictive, and interpretable inference of C. elegans escape dynamics
• Tropical Land Use Land Cover Mapping in Pará (Brazil) using Discriminative Markov Random Fields and Multi-temporal TerraSAR-X Data
• SMASH: Physics-guided Reconstruction of Collisions from Videos
• A Rate-Distortion Framework for Explaining Neural Network Decisions
• Neural-Guided Symbolic Regression with Asymptotic Constraints
• Genetic Architect: Discovering Genomic Structure with Learned Neural Architectures
• Interpretable Parallel Recurrent Neural Networks with Convolutional Attentions for Multi-Modality Activity Modeling
• Interpreting Deep Learning: The Machine Learning Rorschach Test?

Cited by in corpus (93)

• Explaining Deep Neural Networks and Beyond: A Review of Methods and Applications
• Informed Machine Learning -- A Taxonomy and Survey of Integrating Knowledge into Learning Systems
• Social physics
• PhyGeoNet: Physics-Informed Geometry-Adaptive Convolutional Neural Networks for Solving Parameterized Steady-State PDEs on Irregular Domain
• From Anecdotal Evidence to Quantitative Evaluation Methods: A Systematic Review on Evaluating Explainable AI
• On scientific understanding with artificial intelligence
• Autonomous discovery in the chemical sciences part I: Progress
• Autonomous discovery in the chemical sciences part II: Outlook
• Explainable Artificial Intelligence: A Survey of Needs, Techniques, Applications, and Future Direction
• Intelligent metaphotonics empowered by machine learning
• Explainable AI over the Internet of Things (IoT): Overview, State-of-the-Art and Future Directions
• Neural Additive Models: Interpretable Machine Learning with Neural Nets
• Toward Explainable AI for Regression Models
• Beyond Expertise and Roles: A Framework to Characterize the Stakeholders of Interpretable Machine Learning and their Needs
• Elucidating the Behavior of Nanophotonic Structures Through Explainable Machine Learning Algorithms
• Computer-inspired Quantum Experiments
• Towards a Collective Agenda on AI for Earth Science Data Analysis
• Opening the Black-Box: A Systematic Review on Explainable AI in Remote Sensing
• The Road to Explainability is Paved with Bias: Measuring the Fairness of Explanations
• DySMHO: Data-Driven Discovery of Governing Equations for Dynamical Systems via Moving Horizon Optimization
• Deep reinforcement learning for quantum multiparameter estimation
• Conceptual understanding through efficient inverse-design of quantum optical experiments
• Deep Molecular Dreaming: Inverse machine learning for de-novo molecular design and interpretability with surjective representations
• Asset management, condition monitoring and Digital Twins: damage detection and virtual inspection on a reinforced concrete bridge
• Machine learning topological phases in real space
• A deep learning approach to predict the number of k-barriers for intrusion detection over a circular region using wireless sensor networks
• Explainable Artificial Intelligence (XAI) for Malware Analysis: A Survey of Techniques, Applications, and Open Challenges
• The Grammar of Interactive Explanatory Model Analysis
• Comprehensible Artificial Intelligence on Knowledge Graphs: A survey
• Nonperturbative renormalization for the neural network-QFT correspondence
• On the Interpretability of Quantum Neural Networks
• Physically Interpretable Feature Learning and Inverse Design of Supercritical Airfoils
• Explainable Machine Learning with Prior Knowledge: An Overview
• Bridging adaptive management and reinforcement learning for more robust decisions
• Scientific intuition inspired by machine learning generated hypotheses
• Learning quantum dynamics with latent neural ODEs
• Towards falsifiable interpretability research
• 3DMASC: Accessible, explainable 3D point clouds classification. Application to Bi-spectral Topo-bathymetric lidar data
• Scientific Inference With Interpretable Machine Learning: Analyzing Models to Learn About Real-World Phenomena
• Secondary control activation analysed and predicted with explainable AI
• MoËT: Mixture of Expert Trees and its Application to Verifiable Reinforcement Learning
• Physics-informed generative neural network: an application to troposphere temperature prediction
• Comparison of tree-based ensemble algorithms for merging satellite and earth-observed precipitation data at the daily time scale
• Operationally meaningful representations of physical systems in neural networks
• Interpretability in Symbolic Regression: a benchmark of Explanatory Methods using the Feynman data set
• Active learning-assisted neutron spectroscopy with log-Gaussian processes
• What Does Evaluation of Explainable Artificial Intelligence Actually Tell Us? A Case for Compositional and Contextual Validation of XAI Building Blocks
• On gray-box modeling for virtual flow metering
• Deep Q-learning: a robust control approach
• Exploring deterministic frequency deviations with explainable AI
• Disentangled Explanations of Neural Network Predictions by Finding Relevant Subspaces
• Identifiability and physical interpretability of hybrid, gray-box models -- a case study
• Black-box Error Diagnosis in Deep Neural Networks for Computer Vision: a Survey of Tools
• Ensemble learning for blending gridded satellite and gauge-measured precipitation data
• Hydroclimatic time series features at multiple time scales
• Insightful analysis of historical sources at scales beyond human capabilities using unsupervised Machine Learning and XAI
• Features of the Earth's seasonal hydroclimate: Characterizations and comparisons across the Koppen-Geiger climates and across continents
• Selection of powerful radio galaxies with machine learning
• Evolutionary Optimization of Physics-Informed Neural Networks: Evo-PINN Frontiers and Opportunities
• Revealing interactions between HVDC cross-area flows and frequency stability with explainable AI
• Developing Future Human-Centered Smart Cities: Critical Analysis of Smart City Security, Interpretability, and Ethical Challenges
• Emergence of a finite-size-scaling function in the supervised learning of the Ising phase transition
• Interpretable Model-based Hierarchical Reinforcement Learning using Inductive Logic Programming
• Inverse Estimation of Elastic Modulus Using Physics-Informed Generative Adversarial Networks
• A review of machine learning concepts and methods for addressing challenges in probabilistic hydrological post-processing and forecasting
• Revealing drivers and risks for power grid frequency stability with explainable AI
• Practical Trainable Temporal Postprocessor for Multistate Quantum Measurement
• Knowledge-based XAI through CBR: There is more to explanations than models can tell
• Spline parameterization of neural network controls for deep learning
• Mesh-Agnostic Decoders for Supercritical Airfoil Prediction and Inverse Design
• Multi-Exit Kolmogorov-Arnold Networks: enhancing accuracy and parsimony
• Comparison of generalised additive models and neural networks in applications: A systematic review
• Benchmark datasets for biomedical knowledge graphs with negative statements
• On Explainability in AI-Solutions: A Cross-Domain Survey
• On Locality of Local Explanation Models
• A Causal Lens for Peeking into Black Box Predictive Models: Predictive Model Interpretation via Causal Attribution
• Learning to Make Chemical Predictions: the Interplay of Feature Representation, Data, and Machine Learning Algorithms
• Uncertainties of Satellite-based Essential Climate Variables from Deep Learning
• Data-Driven Theory-guided Learning of Partial Differential Equations using SimultaNeous Basis Function Approximation and Parameter Estimation (SNAPE)
• SRoll3: A neural network approach to reduce large-scale systematic effects in the Planck High Frequency Instrument maps
• Interpretability of machine-learning models in physical sciences
• Deep-Learning Discovers Macroscopic Governing Equations for Viscous Gravity Currents from Microscopic Simulation Data
• Explainable Machine Learning in Healthcare: Methods, Interpretation, and Applications for Clinical Research
• A Clustering-aided Ensemble Method for Predicting Ridesourcing Demand in Chicago
• Sub-Setting Algorithm for Training Data Selection in Pattern Recognition
• Discovery of the Hidden State in Ionic Models Using a Domain-Specific Recurrent Neural Network
• Explanation from Specification
• Molecular origin of blood-based infrared fingerprints
• SMT-EX: An Explainable Surrogate Modeling Toolbox for Mixed-Variables Design Exploration
• Principle of learning sign rules by neural networks in qubit lattice models
• On the Use of Interpretable Machine Learning for the Management of Data Quality
• Time series features for supporting hydrometeorological explorations and predictions in ungauged locations using large datasets
• Toward Building Science Discovery Machines