Generative Methods In Quantum Computing For First Principle Simulations

Date

2025-8-26

Author

Korhan, Hamza

Metadata

Show full item record

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Item Usage Stats

1275
views

0
downloads

Cite This

This thesis introduces GPT-QE, a quantum-inspired total energy method that addresses the ansatz design limitations of the Variational Quantum Eigensolver (VQE). The framework's key contribution is an iterative online learning loop where a generative Transformer model improves itself by periodically retraining on a growing dataset of high-quality circuits it generates from a chemically-motivated, native UCCSD operator pool. Implemented using PennyLane and PyTorch, the framework is benchmarked by computing the potential energy surface of the H4 molecule, demonstrating its ability to accurately reproduce ground state energies in close agreement with Full Configuration Interaction (FCI) results.

Subject Keywords

Quantum Chemistry, Generative Models, Quantum Computing, Variational Quantum Eigensolver, Kuantum Kimyası, Üretken Modeller, Kuantum Hesaplama, Anahtar Varyasyonel Kuantum Özdeğer Çözücü

URI

https://hdl.handle.net/11511/115533

Collections

Graduate School of Natural and Applied Sciences, Thesis