Research on methods of simulation of Twirled Readout Error eXtinction (TREX)

[ikkoham]

Description

Twirled Readout Error eXtinction (TREX) is a readout error mitigation proposed in https://doi.org/10.1103/PhysRevA.105.032620
This project looks for methods to simulate this technique at high speed without running calibration directly.

Deliverables

Research Paper or Pull Request to Qiskit Aer

Mentors details

• Mentor 1
  • Name:Ikko Hamamura
  • GitHub ID: ikkoham
  • What they do: Qiskit core developer and Researcher of IBM Quantum

Number of mentees

1

Type of mentees

• Mentee 1
  • Required:
    • Deep understanding the paper.
  • Nice to have:
    • Python, Qiskit

[q-inho]

I am interested in this project! :)

[deveshq]

I'm interested in this project.

[RagaviKrishnamoorthy]

I find the topic of addressing readout errors in quantum systems fascinating and believe that it has significant implications for the development of quantum technologies. The fact that this method can be applied to general noise models without making any assumptions about the specific form of the noise is particularly intriguing. I am excited to learn more about the bounds on the error in the expectation value and how they relate to the sample complexity. I am very interested in this project and would love to contribute in any way possible.

[q-inho]

Checkpoint 2:

The project focuses on finding ways to simulate TREX technology at high speeds. To do this, we loaded the account of IBM and chose a backend called 'ibm_lagos'. To analyze the properties of the backend, we wrote a function describing the properties of each qubit. This function outputs the T1 time, T2 time, U2 gate error, U2 gate duration, resonant frequency, etc. of the qubit.

Next, we calculated the A matrix for each qubit. This matrix represents the measurement error of the qubit, which was converted into a sparse matrix form and stored. The sparse A matrix thus created is output for each qubit.

We also wrote a function that calculates the A* matrix. This matrix is a variant of the A matrix, which is used to correct measurement errors in qubits. This matrix was also converted into a sparse matrix form and stored, and was output for each qubit.

We performed the task of converting this calculated A* matrix into a probability matrix. This probability matrix was used to generate a ReadoutError. This ReadoutError was used to generate a noise model, which was set in the AerSimulator.

The RealAmplitude circuit was simulated using the simulator that set up using our Noise model. The simulation results are displayed as histograms.

We also performed simulation of the circuit in the presence of our NoiseModel. This result was also printed to show the results with noisy simulation.

The next step in this project is to fine-tune the noise model and run the simulation using various circuits and parameters. Also, we will research and study to have a deeper understanding and implement the high-speed simulation of TREX technology.