A Strategic Approach to Machine Learning for Material Science: How to Tackle Real-World Challenges and Avoid Pitfalls

Abstract

The exponential growth and success of machine learning (ML) has resulted in its application in all scientific domains including material science. Advancement in experimental techniques has led to an increase in the volume of material science data encouraging material scientists to investigate data-driven solutions to scientific problems. While the resources available to get started with ML are ever increasing, there is little literature on traversing through the space of decisions that need to be made to implement a robust and trustworthy ML solution. A lack of such resources leads to researchers wading through articles and papers trying to determine the best approach for their problem and sometimes also falling prey to pitfalls in a real-world scenario. This paper aims to act as a guide for researchers who want to strategically approach a ML solution to their problem through the use of domain knowledge and systematic evaluation of the major aspects of a ML pipeline. We focus on four aspects of the ML pipeline: (1) problem formulation, (2) data curation, (3) feature representation and model selection, and (4) model generalizability and real-world performance. In each case, we discuss the space of decisions, provide examples from scientific literature, and illustrate how different choices can affect the outcome through a case study of predicting compressive strength of uniaxially pressed molecular solid, 2,4,6-triamino-1,3,5-trinitrobenzene (TATB) samples. Using a similar approach of critical thinking along with rigorous evaluation and diagnostics, researchers can be assured of the reliability of predictions from their ML models.

Deep Generative Modeling of Infrared Images Provides Signature of Cracking in Cross-Linked Polyethylene Pipe

“…Machine learning and deep learning approaches are increasingly being applied to spectroscopy data and material characterization. − We have previously reported on a deep learning approach that we applied to the study of complex IR spectroscopy data, in which we trained an artificial neural network β-variational autoencoder (β-VAE) on a data set of 25,000 IR spectra of PEX-a pipes . β-VAEs are deep generative models capable of learning disentangled (independent and interpretable) representations of the generative factors responsible for variance in data. ,, For PEX-a pipe IR spectra, the physicochemical processes that can occur during pipe aging, degradation, and cracking are the underlying generative factors responsible for the variance in the spectra .…”

Quantitative Structure–Retention Relationship Analysis of Polycyclic Aromatic Compounds in Ultra-High Performance Chromatography

“…During the training process, the model is evaluated on the validation set at regular intervals to monitor its performance and detect any signs of overfitting. If the model’s performance on the validation set starts to degrade while its performance on the training set continues to improve, this is a sign of overfitting, and the training process can be stopped [ 27 ]. The slightest validation error is indeed a good criterion to stop training the network.…”

“…Quantitative structure–retention relationships (QSRRs) represent the theoretical description of chromatographic retention behavior using physicochemical properties derived from the chemical structure of analytes and from the effect of chromatographic conditions [ 21 , 22 , 23 , 24 , 25 , 26 , 27 ]. A method of optimization to represent the correct geometry of each analyte is required to provide data for the calculations of molecular descriptors.…”

Quantum Machine Learning in Materials Prediction: A Case Study on ABO3 Perovskite Structures

“…It has already contributed significantly to the prediction of the properties of materials and accelerated the development of molecular systems with specified properties. Machine learning algorithms are capable of analyzing complex data sets and uncovering hidden patterns that have greatly assisted researchers in understanding and designing new materials . The integration of experimental data with ML techniques has the potential to revolutionize the field of materials science, enabling scientists to make informed decisions and drive innovation in materials research.…”