Quantum Circuits, Feature Maps, and Expanded Pseudo-Entropy: A Categorical Theoretic Analysis of Encoding Real-World Data into a Quantum Computer (2410.22084v2)

Abstract: This manuscripts proposes a new and novel numerical method to the determine the efficacy of an encoding scheme to map real-world data into a quantum circuit. The method calculates the Shannon entropy of each of the data points from a point-cloud, hence, samples from an embedded manifold, and calculates the expanded concept of pseudo-entropy applied to each respective quantum operator that comes from a given quantum feature map, and not the density operator. In the recent decade, there has been a continuous advancement of translating machine learning into a quantum circuit with many promising results. For quantum machine learning, a major underlying question is how to encode real-world data into a quantum circuit without losing information and adding noise. A few notable methods derived are expressibility, where the distribution of the output of states from the circuit are compared against the Haar probability measure with information theoretic techniques, and expressivity, a method that maps the expectation of a quantum circuit to the space of complex functions via a partial Fourier series, noting that more intricate the function the more expressive, and using the symmetry embedded within the data to derive a quantum feature map. The proposed pseudo-entropy method is discussed to and empirically shown to generalize these methods. Furthermore, this method is argued to also generalize symmetric quantum feature maps. The discussions and arguments are a reasonable basis for understanding the connections but require deeper mathematical analysis.