INQNET seminars

Enhancing Quantum Transducers using Deep Reinforcement Learning

by Dr Mekena Metcalf (Lawrence Berkeley National Lab)


Interconnection of qubits contained in different dilution refrigerators using fiber optic cables requires the conversion of static qubits to optical, flying qubits. This process, known as quantum transduction, is difficult since the energy conversion requires strong coupling between the microwave and optical cavities. The optomechanical transducer has demonstrated the highest conversion efficiency to date. Artificial intelligence can enhance this efficiency by bridging experimental data and theoretical models. We employ Deep Reinforcement Learning to optimize the microwave to optical conversion efficiency of an optomechanical quantum transducer. We frame a quantum control problem using the cavity drive pulses for the optomechanical Hamiltonian. The AI agent performs actions on the control pulses to maximize the number of microwave photons converted to optical photons in the presence of dissipation and thermal noise. Our AI agent is trained using a device simulator that models the optomechanical experimental design developed by the Regal group at NIST.


Topic: INQNET seminar 
Time: March 22, 2021 12:30 PM Pacific Time (US and Canada)

Join Zoom Meeting

Meeting ID: 933 0458 4361

One tap mobile
+16699006833,,93304584361# US (San Jose)
+13462487799,,93304584361# US (Houston)

Organized by

Nikolai Lauk