Quantum gate control of polar molecules with machine learning.

Area of Science:

• Quantum Information Science
• Molecular Physics
• Computational Chemistry

Background:

• Quantum computing requires high-fidelity quantum gates for reliable operations.
• Ultracold polar molecules offer a promising platform for building qubits due to their controllable interactions.
• Efficient control strategies are crucial for realizing molecular quantum computing.

Purpose of the Study:

• To propose and demonstrate a scheme for achieving basic quantum gates using ultracold polar molecules.
• To utilize deep reinforcement learning (DRL) for optimizing control pulse sequences for molecular qubits.
• To investigate the feasibility of molecule-based quantum computing through simulation of quantum circuits.

Main Methods:

• Encoding qubits in Ytterbium Fluoride (YbF) molecules trapped in an electric field gradient.
• Coupling molecular qubits via dipole-dipole interactions.
• Applying time-dependent control pulses and optimizing them using a Markov decision process framework with deep reinforcement learning (DRL).

Main Results:

• Discovered optimal control pulse sequences for NOT, controlled NOT, and Hadamard two-qubit gates with high fidelities.
• Analyzed the population dynamics of YbF molecules under the discovered gate sequences.
• Successfully simulated a quantum circuit for entanglement using the optimized gate sequences.

Conclusions:

• Deep reinforcement learning (DRL) provides an effective method for designing high-fidelity quantum gates for molecular qubits.
• Ultracold polar molecules in pendular states are a viable platform for practical molecule-based quantum computing.
• The developed control schemes offer new insights into efficient manipulation of molecular systems for quantum information processing.