Development of a Recommender System for Design of Next Generation Battery Materials



The design space for the next generation of energy generation and storage materials is huge.  With billions of possible combinations of elements it is no longer possible to test them all, due the prohibitive cost of raw materials and the time and expense involved in synthesis and characterisation.  Scientists need guidance as to which combinations to try, based on the collectively knowledge in the scientific literature, but this has also become too vast and complicated to be used easily navigated.  This is not dissimilar to the huge array of products consumers are confronted which when deciding what to buy on global e-commerce platforms such as Amazon.  Amazon simplifies this decision for customers using recommender systems, and so there is compelling case for the same approach to guide scientists and engineers in their similar decision making processes.  However, the use of recommender systems to guide scientific research is largely unexplored.  In this project you will develop a recommender system for a data set of energy capable of suggesting which materials a scientist or engineer should make to meet their required performance criteria, and eliminate the need for costly, unguided trial-and-error experimentation. A case study data set will be provided.


Develop a recommender system capable of predicting the chemical formula of a new energy material, based on a set of performance criteria


Python programming and experience in data science and machine learning is essential (such as COMP3720, COMP4660, COMP4670, COMP6670, COMP8420).  Familiarity with platforms such as scikit-learn, Pytorch, Tensorflow and Keras is desirable.


This is a 24cp project


machine learning, materials informatics, data science, recommender systems

Updated:  10 August 2021/Responsible Officer:  Head of School/Page Contact:  CECS Webmaster