Theoretical and computational studies are integral parts of research in interdisciplinary areas of science and technology. The advent of powerful modern computers, developments in sophisticated algorithms and theories, and access to a large amount of data from previous studies suggest that in the future, computational techniques would continue to play a dominant role in both fundamental and applied research. However, currently used computational methods have well-known limitations. Although a few groups have introduced automated reaction search algorithms and high-throughput studies with some success, myriads of unique possible pathways and combinations should be investigated by using accurate theoretical methods to furnish a reliable theoretical prediction of the reaction outcome. This makes the calculations prohibitively expensive, highly time consuming, and tedious compared to the actual experiments.

Inspiring from the recent success of deep-learning and artificial neural networks, we propose to apply them for the designing of novel materials for energy related applications. We would like to apply the principles of machine leaning to design solar energy materials, batteries, and energy storage devices. We would use existing machine learning algorithms and also develop our own code to tackle with the challenging problems in chemistry and materials science. A combination of fields would help us to analyze, understand, and rationalize the structure-activity relationships of numerous candidate systems and select the optimum ones for the experimental realization.