Gravitational waves carry information about far away phenomena which astronomers can use to form a better picture of the universe. Over the past few years, numerous gravitational waves have been detected using a method called matched filtering; however, this approach has two main drawbacks. First, it requires knowledge of the sought-after waveform beforehand. Second, it is infeasible for gravitational wave searches which are precise or cover a long time span. While attempts have been made to use classical machine learning to overcome these difficulties, the usage of quantum machine learning in this context is relatively unexplored. Quantum variational rewinding is a hybrid quantum-classical machine learning algorithm for time series anomaly detection. This work makes the first steps in investigating its potential for gravitational wave detection. Training and testing the algorithm was done without depending on synthetic data. In a noiseless simulated environment, quantum variational rewinding was able to pick out all confirmed gravitational wave signals from the background noise of a detector with perfect accuracy when filtering was applied before hand and was able to do so to a high degree of success without filtering. Anomaly detection with this algorithm was performed with two signals on quantum hardware as a proof of concept. Using Fisher information, both the capacity and the trainability of quantum variational rewinding are described so that future approaches can be compared in these regards. With new gravitational wave detectors under development, even more data will be generated for analysis in the future, compounding the existing bottlenecks in the search pipeline. As such, approaches which help tackle issues with the current approach to gravitational wave analysis are in demand, and the empirical results from this work assert that quantum computing should not be overlooked in this process.

Evaluating the performance of quantum devices is an important step towards scaling quantum devices and eventually using them in practice. The great number of available quantum metrics and the different hardware technologies used to develop quantum computers complicate this evaluation. In addition, different computational paradigms implement quantum operations in different ways. We add to the landscape of quantum metrics by extending the Q-score metric of Atos to the Q-score Max-Clique. To our knowledge, this yields the first application-level metric which allows comparison of three different paradigms of quantum computing. This metric is evaluated on these computational quantum paradigms -- quantum annealing, gate-based quantum computing, and photonic quantum computing -- and the results are compared to those obtained by classical solvers.