Presented by Aroosa Ijaz, U of Waterloo
Linear algebraic methods for machine learning, such as kernel methods, operate by embedding data points as vectors in a high dimensional vector space, and by applying methods of linear algebra to classify and to discriminate between embedded clusters of data. Quantum mechanics represents a natural setting for enacting such linear algebraic machine learning methods. When data points are embedded as quantum states, we naturally end up using vectors that live in high-dimensional vector spaces. In this session, we will talk about the power of quantum embeddings for machine learning and how the way we encode data might carry a lot of weight in the performance of quantum learning models. As an example, we will see that the performance of quantum classifiers can be completely determined by the quantum feature map that performs the embedding.