Abstract
Recent advances in classical machine learning have shown that creating modelswith inductive biases encoding the symmetries of a problem can greatly improveperformance. Importation of these ideas, combined with an existing rich body ofwork at the nexus of quantum theory and symmetry, has given rise to the fieldof Geometric Quantum Machine Learning (GQML). Following the success of itsclassical counterpart, it is reasonable to expect that GQML will play a crucialrole in developing problem-specific and quantum-aware models capable ofachieving a computational advantage. Despite the simplicity of the main idea ofGQML -- create architectures respecting the symmetries of the data -- itspractical implementation requires a significant amount of knowledge of grouprepresentation theory. We present an introduction to representation theorytools from the optics of quantum learning, driven by key examples involvingdiscrete and continuous groups. These examples are sewn together by anexposition outlining the formal capture of GQML symmetries via "labelinvariance under the action of a group representation", a brief (but rigorous)tour through finite and compact Lie group representation theory, areexamination of ubiquitous tools like Haar integration and twirling, and anoverview of some successful strategies for detecting symmetries.