Everyone is welcome to attend the online QuSoft seminar this week with David Sutter on 'The power of quantum neural networks'.
Abstract:
It is unknown if near-term quantum computers are advantageous for machine learning tasks. In this talk, we address this question by trying to understand how powerful and trainable quantum machine learning models are, relative to popular classical neural networks. We propose the effective dimension — a novel measure that captures these qualities — and prove that it can be used to assess any statistical model's ability to generalize on new data. Intuitively, the effective dimension tells us how useful the parameters in a model are and since it depends on the Fisher information we can also gauge how well a model trains. We demonstrate numerically that a class of quantum neural networks is able to achieve a significantly better effective dimension than comparable feedforward networks and train faster, suggesting an advantage for quantum machine learning, which we verify on real quantum hardware.
This is joint work with Amira Abbas, Christa Zoufal, Aurelien Lucchi, Alessio Figalli, Stefan Woerner and is based on
https://arxiv.org/abs/2011.00027
Please contact Subhasree Patro or Jop Briet if you like to join.
