Abstract by Kennedy Lincoln
Physics and Astronomy
Representations in Machine Learning
Currently, computationally finding new materials with high accuracy is expensive and time-consuming. Machine learning can help us reduce the cost of calculations to a great extent. For machine learning methods to be effective, we need a unique representation of crystal structures. Describing the atomic systems require a representation that is invariant to translations, rotations, and permutations of the crystal structure. In this talk, we will present the importance of representations in machine learning with a simple example.