![Getting started
with Qiskit
IBM Quantum 3
pip install 'qiskit[visualization]'
pip install qiskit-ibm-runtime
To access IBM hardware, set up your credentials
through one of:
• IBM Quantum Platform
• IBM Cloud
https://docs.quantum-computing.ibm.com/start](https://image.slidesharecdn.com/qiskit-101-250720214913-4b702ad9/85/introduction-to-ibm-qiskit-101-to-learn-quantum-3-320.jpg)
![Configuring runtime transpilation
The primitives expose preset
transpilation configurations via
the optimization_level
option.
You can also elect to
disable runtime transpilation
using the
skip_transpilation
option.
IBM Quantum 29
Optimization level Effect
0
No optimization: typically used for
hardware characterization or debugging
• Basis translation
• Layout (as specified)
• Routing (stochastic swaps)
1
[Default]
Light optimization:
• Layout (trivial → vf2 → SabreLayout if
routing is required)
• Routing (SabreSWAPs if needed)
• 1Q gate optimization
• Error suppression: dynamical decoupling](https://image.slidesharecdn.com/qiskit-101-250720214913-4b702ad9/85/introduction-to-ibm-qiskit-101-to-learn-quantum-29-320.jpg)
![Configuring error mitigation
The primitives expose error
mitigation via the
resilience_level option.
Higher resilience levels incur a
greater cost.
IBM Quantum 30
Resilience level Definition Estimator Sampler
0 No mitigation
1
[Default]
Minimal mitigation costs:
Mitigate error associated
with readout errors
Twirled Readout Error
eXtinction (TREX)
Matrix-free Measurement
Mitigation (M3)
2
Medium mitigation costs.
Typically reduces bias in
estimators, but is not
guaranteed to be zero-
bias.
Zero Noise Extrapolation
(ZNE)
3
Heavy mitigation with
layer sampling.
Theoretically expected to
deliver zero-bias
estimators.
Probabilistic Error
Cancellation (PEC)](https://image.slidesharecdn.com/qiskit-101-250720214913-4b702ad9/85/introduction-to-ibm-qiskit-101-to-learn-quantum-30-320.jpg)
introduction to ibm qiskit-101 to learn quantum
- 2. What is Qiskit? Qiskit is an open-source SDK for working with quantum computers at the level of extended quantum circuits, operators, and primitives. https://github.com/Qiskit/qiskit https://docs.quantum-computing.ibm.com/ The Qiskit Ecosystem is a collection of software and tutorials that build on or extend Qiskit. http://qiskit.github.io/ecosystem IBM Quantum 2
- 3. Getting started with Qiskit IBM Quantum 3 pip install 'qiskit[visualization]' pip install qiskit-ibm-runtime To access IBM hardware, set up your credentials through one of: • IBM Quantum Platform • IBM Cloud https://docs.quantum-computing.ibm.com/start
- 4. Qiskit workflow Build Design and develop quantum circuits https://docs.quantum.ibm.com/build Transpile Optimize circuits to efficiently run on hardware https://docs.quantum.ibm.com/transpile Verify Validate and evaluate your quantum circuits https://docs.quantum.ibm.com/verify Run Execute programs on quantum hardware via Qiskit Runtime https://docs.quantum.ibm.com/run IBM Quantum 4
- 5. Build IBM Quantum 5 In the build phase, you create quantum programs that represent the problem you are solving.
- 6. Build a circuit with Qiskit The foundation of quantum programs are quantum circuits, which consist of operations - including gates, measurement, and reset - that manipulate qubits in the quantum computer. IBM Quantum 6 To build a circuit: • Initialize a register of qubits • Add the qubits to a circuit • Perform operations on those qubits
- 7. The output of a quantum circuit is encapsulated by primitives Sampler primitive Output is a (quasi-)probability distribution over possible measurement outcomes Circuit should include measurements Estimator primitive Output is the expectation value of an observable (e.g. cost function of an optimization problem) Circuit should not include measurements IBM Quantum 7
- 8. Qiskit includes a library of standard gates and common circuits Standard gates Standard quantum gates such as Hadamard, Pauli rotation gates, CNOT, Quantum Fourier Transform... Variational ansatzes Parameterized quantum circuits for chemistry and combinatorial optimization, include hardware efficient ansatzes Benchmarking circuits Circuits used for measuring quantum volume and other benchmarks of quantum devices And more! IBM Quantum 8
- 9. Transpile IBM Quantum 9 In the transpile phase, you rewrite your circuit into a form that can be efficiently executed on a quantum device
- 10. Real quantum devices are subject to constraints Basis gate set Only a limited set of gates can be executed directly on the hardware. Other gates must be rewritten in terms of these basis gates. Qubit connectivity Only certain pairs of qubits can be directly interacted with each other. Errors Each operation has a chance of error, so circuit optimizations can greatly affect performance. 10 IBM Quantum
- 11. Transpiler definitions Transpilation occurs in a series of passes, with the output of one pass becoming the input to the next. A transpiler pass defines a standalone circuit transformation. A pass manager is a list of transpiler passes grouped into a logical unit. A staged pass manager is a list of pass managers, with each one representing a discrete stage of a transpilation pipeline. IBM Quantum 11
- 12. Transpiler stages 1 Initialization The circuit is prepared for transpilation. 2 Layout The abstract qubits of the circuit are mapped to physical qubits on the device. 3 Routing Swap gates are inserted to enable interactions between qubits that are not physically connected. 4 Translation The gates of the circuit are translated to the basis gate set of the device. 5 Optimization The circuit is rewritten to minimize its depth or number of operations in order to decrease the effect of errors. 6 Scheduling The precise timing of pulse sequences is determined based on hardware information. IBM Quantum 12
- 13. Transpile a circuit Qiskit can transpile your circuit for you with reasonable default settings. IBM Quantum 13 To transpile a circuit using Qiskit's presets: • Choose which device backend you want to target • Create a preset staged pass manager with your desired optimization level • Run the pass manager on the circuit
- 14. You don't need to transpile your circuits to submit them to Qiskit Runtime! But for the best performance, you should design your own transpilation pipeline. If you transpile your circuits locally, remember to set skip_transpilation=True (more on this later). IBM Quantum 14
- 15. Verify IBM Quantum 15 In the verify phase, you test your quantum programs by running them on simulated devices and exploring their performance under realistic device noise models.
- 16. The cost of simulating quantum circuits scales exponentially with the number of qubits. Test smaller versions of your circuit Modify the circuit so that it becomes classically simulable (e.g. a stabilizer circuit) IBM Quantum 16 Simulating quantum circuits Roughly speaking, circuits larger than 50 qubits can't be simulated. For such large circuits, you can:
- 17. Several simulation tools are available for verifying your quantum programs Reference primitives included with Qiskit For exact simulation of small quantum circuits Qiskit Aer primitives For higher-performance simulation that can handle larger circuits, or to incorporate noise models https://qiskit.org/ecosystem/aer/ Qiskit Aer stabilizer simulator For efficient simulation of stabilizer circuits 17 IBM Quantum Simulation tools
- 18. Use the Estimator primitive from Qiskit Aer The Estimator primitive is used to estimate the expectation value of an observable. IBM Quantum 18 To use the Estimator primitive, pass it: • A circuit without measurements • An observable Measure expectation value of ZZ
- 19. Simulate noise Real quantum devices are subject to errors. These errors vary across different devices, different qubits and connections on the same device, and time. In the near term, a variety of error-mitigation techniques help us counter the effects of noise. In the long term, we will use error-correction to build fault- tolerant quantum computers. IBM Quantum 19 Calibration data for ibm_brisbane 9am ET, Nov 16, 2023 https://quantum-computing.ibm.com/
- 20. Build noise models with Qiskit Aer You can use Qiskit Aer to build and simulate noise models. IBM Quantum 20 You can initialize a noise model with parameters set from the calibration data of a real backend. You can also build your own custom noise models from scratch. Depolarizing error:
- 21. Stabilizer circuits can be simulated efficiently Stabilizer circuits are a special restricted class of circuits that can be efficiently simulated classically. IBM Quantum 21 To simulate a stabilizer circuit, use a Qiskit Aer primitive with the "stabilizer" simulation method. Stabilizer circuit if every 𝜃 is an integer multiple of 𝜋/2
- 22. Run IBM Quantum 22 In the run phase, you send your quantum program to be executed on a quantum system.
- 23. Anatomy of a quantum computing service IBM Quantum 23 User Cloud service (Qiskit Runtime) Server in laboratory Control electronics Quantum processor Quantum programming language Pulse sequences
- 24. Run a circuit on quantum hardware using Qiskit Runtime Running a circuit on a real quantum device is similar to running it on a simulator. The main difference is that you use a Qiskit Runtime primitive and choose the hardware backend you want to use. IBM Quantum 24 To run a circuit on quantum hardware: • Initialize the Qiskit Runtime service (requires setting up account credentials beforehand) • Initialize a Qiskit Runtime primitive • Invoke the primitive with your circuit
- 25. Qiskit Runtime sessions A session allows a collection of jobs to be grouped and jointly scheduled by the Qiskit Runtime service, facilitating iterative use of quantum computers without incurring queuing delays on each iteration. Jobs within an active session take priority over other queued jobs. A session becomes active when its first job starts running. A session stays active until one of the following happens: • Its maximum timeout value is reached. • Its interactive timeout value is reached. In this case the session is deactivated but can be resumed if another session job starts running. • The session is closed or cancelled. Note: The queuing time does not decrease for the first job submitted within a session. IBM Quantum 25
- 26. IBM Quantum 26 Qiskit Runtime sessions
- 27. Run jobs in a session A session is typically created as a context manager using the with statement. The context manager takes care of closing the session when you're done. IBM Quantum 27 To run jobs in a session: • Initialize the Qiskit Runtime service • Initialize a session as a context manager, specifying the backend • Initialize a Qiskit Runtime primitive within the session • Invoke the primitive as usual
- 28. Several options are available to customize the behavior of Qiskit Runtime primitives Shots The number of measurement shots used by the primitives to compute their results. Increasing this value lowers statistical error, but increases running time. Runtime transpilation The primitives may perform runtime transpilation to optimize circuits, with the degree of optimization controlled by an optimization level option. The optimization level you choose affects the transpilation strategy, with higher levels invoking more expensive or aggressive optimizations. Error mitigation Several error mitigation techniques are available to help you counter the effects of device noise. These techniques incur computational overhead and should be evaluated case-by-case. 28 IBM Quantum Customize Qiskit Runtime behavior
- 29. Configuring runtime transpilation The primitives expose preset transpilation configurations via the optimization_level option. You can also elect to disable runtime transpilation using the skip_transpilation option. IBM Quantum 29 Optimization level Effect 0 No optimization: typically used for hardware characterization or debugging • Basis translation • Layout (as specified) • Routing (stochastic swaps) 1 [Default] Light optimization: • Layout (trivial → vf2 → SabreLayout if routing is required) • Routing (SabreSWAPs if needed) • 1Q gate optimization • Error suppression: dynamical decoupling
- 30. Configuring error mitigation The primitives expose error mitigation via the resilience_level option. Higher resilience levels incur a greater cost. IBM Quantum 30 Resilience level Definition Estimator Sampler 0 No mitigation 1 [Default] Minimal mitigation costs: Mitigate error associated with readout errors Twirled Readout Error eXtinction (TREX) Matrix-free Measurement Mitigation (M3) 2 Medium mitigation costs. Typically reduces bias in estimators, but is not guaranteed to be zero- bias. Zero Noise Extrapolation (ZNE) 3 Heavy mitigation with layer sampling. Theoretically expected to deliver zero-bias estimators. Probabilistic Error Cancellation (PEC)
- 31. Links IBM Quantum Documentation https://docs.quantum-computing.ibm.com/ Access the documentation for Qiskit and IBM Quantum services. IBM Quantum Learning https://learning.quantum-computing.ibm.com/ Learn the basics of quantum computing, and how to use IBM Quantum services and systems to solve real-world problems. Qiskit YouTube https://www.youtube.com/qiskit Join us for engaging lectures, tips & tricks, tutorials, community updates and access to exclusive Qiskit content! IBM Quantum 31