Abstract
The present disclosure relates to a system and method for Hamiltonian-Based Quantum Generative Autoencoder. The autoencoder leverages a novel Hamiltonian based quantum Al generative model to resolve anomaly detection in financial transaction data. An hybrid architecture, i.e., potential of quantum computing and generative AI based approach is leveraged in this disclosure to capture complex patterns and correlations in high-dimensional financial data. This system employes a series of parameterized Hamiltonians at least for processes like encoding, latent space representation, and decoding, which are implemented as quantum circuits using Trotterization techniques. Particularly, the system incorporates physical-inspired principles into the area of financial data analysis, using quantum evolution under carefully constructed Hamiltonians to perform data compression and reconstruction. The encoding process maps transaction data onto quantum states, which is then compressed into a lower-dimensional latent space. The decoding process reconstructs the input data received from the latent space and enables the generation of new, synthetic transaction data, The encoding process in this disclosure may be viewed as a quantum channel that maximizes the mutual information between the input and the latent space, while the decoding process aims to minimize the quantum relative entropy between the original and reconstructed States.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
Guo, Fuming, "HAMILTONIAN-BASED QUANTUM GENERATIVE AUTOENCODER SYSTEM AND METHOD THEREOF", Technical Disclosure Commons, ()
https://www.tdcommons.org/dpubs_series/8587