Noise-assisted variational quantum thermalization.

Area of Science:

• Quantum Computing
• Quantum Information Science
• Condensed Matter Physics

Background:

• Preparing thermal states is crucial for quantum simulations and machine learning.
• Existing variational methods face challenges with scalable cost functions, purification, and noise mitigation.

Purpose of the Study:

• To propose a new algorithm for thermal state preparation that leverages quantum circuit noise.
• To address scalability, purification, and noise challenges in near-term quantum devices.

Main Methods:

• Developed a variational quantum circuit architecture with controlled depolarizing noise.
• Derived a closed-form approximation of free energy as a cost function.
• Evaluated the algorithm on diverse Hamiltonians and system sizes.

Main Results:

• Achieved high fidelity thermal state approximation for several systems.
• Demonstrated that algorithm performance is temperature-dependent, excelling at high/low temperatures but facing challenges in intermediate ranges.
• Identified a specific temperature range where learning the thermal state is more difficult.

Conclusions:

• Noise-assisted thermal state preparation is a viable approach for quantum computing.
• The proposed method offers a promising direction for overcoming limitations in variational quantum algorithms.
• Further research into exploiting noise in quantum algorithms is warranted.