(Nanowerk News) Advanced materials are urgently needed for everyday life, be it in high technology, mobility, infrastructure, green energy or medicine. However, traditional ways of discovering and exploring new materials face limitations due to the complexity of the chemical composition, structure, and properties targeted. In addition, new materials should not only enable new applications, but also include sustainable ways of production, use and recycling.
Researchers from the Max-Planck-Institut für Eisenforschung (MPIE) review the state of physics-based modeling and discuss how combining this approach with artificial intelligence can open up so far untapped spaces for the design of complex materials.
They publish their perspectives in journals Natural Computational Science (“Accelerating the design of materials with complex compositions through physics-based artificial intelligence”).
Combining a physics-based approach with artificial intelligence
In order to meet the demands of technological and environmental challenges, increasingly demanding and multiplying material properties must be considered, thereby making alloys more complex in terms of composition, synthesis, processing and recycling. Changes in these parameters require changes in the microstructure, which directly affects the material properties. This complexity needs to be understood to allow predicting the structure and properties of the material. The computational material design approach plays an important role here.
“The way we design new materials today relies exclusively on physics-based simulations and experiments. This approach can suffer from certain limitations when it comes to the quantitative prediction of the equilibrium of high-dimensional phases and in particular to the microstructure and resulting non-equilibrium properties. In addition, many models related to microstructure and properties use simplified approximations and depend on a large number of variables. However, the question remains whether and how these degrees of freedom still account for the complexity of the material”, explained Professor Dierk Raabe, director of MPIE and first author of this publication.
This paper compares physics-based simulations, such as molecular dynamics and ab initio simulations with descriptor-based modeling and advanced artificial intelligence approaches. While physics-based simulations are often too expensive to predict materials with complex compositions, the use of artificial intelligence (AI) has several advantages.
“AI is able to automatically extract thermodynamic and microstructural features from large data sets obtained from electronic, atomistic and continuum simulations with high predictive power”, says Professor Jörg Neugebauer, director at MPIE and one of the authors of the publication.
Enhance machine learning with large data sets
Since the predictive power of artificial intelligence depends on the availability of large data sets, a way is needed to overcome this barrier. One possibility is to use active learning cycles, in which a machine learning model is trained with an initially small subset of labeled data. The model predictions are then filtered by the labeling unit which feeds the high-quality data back into the labeled record set and the machine learning model is run again. This step-by-step approach results in a high-quality final data set that can be used for accurate predictions.
There are still many open questions for the use of artificial intelligence in materials science: how to deal with sparse and noisy data? How to consider interesting outliers or ‘mismatches’? How to implement unwanted element intrusion from synthesis or recycling? However, when it comes to designing complex alloy compositions, artificial intelligence will play a more important role in the near future, especially with the development of algorithms, and the availability of high-quality material datasets and high-performance computing resources.