Hybrid Algorithms
28-4-23
Syllabus proposed
Module 1:
Fundamentals of Quantum mechanics, Vector Algebra, Introduction to Dirac notations
Introduction to Qubits, Quantum parallelism, Entanglement
Module 2:
EPR’s paradox, Bell’s inequality, Introduction to Quantum Cryptography
Quantum gates and circuits
Module 3:
Quantum algorithms – D-J algorithm, B-V algorithm, Shor’s factorization problem, Grover’s search algorithm
Introduction to Quantum machine learning schemes
Module 4:
Quantum error correction and mitigation schemes
Introduction to different Quantum hardware
Hands -on experience on real quantum hardware and quantum emulators
Certificate from MU and Tech Mahindra in association with AWS
Hybrid Quantum-Classical Algorithms
Classical computers are good at tasks - performing arithmetic, data storage,
feature extraction, etc.
Quantum computers are good at - interfering the complex data into higher
dimensional space (not possible in classical computers).
Hybrid models pave way to get the best of both worlds.
Data
Classical- Classical-
Classical Quantum
Algorithm
Quantum- Quantum-
Classical Quantum
Hybrid Quantum Algorithms - Generalized frame work
Manuela Weigold, Johanna Barzen, Frank Leymann, and Marie Salm. 2022. Data encoding patterns for quantum computing. In Proceedings of the 27th Conference on Pattern Languages of Programs (PLoP '20). The Hillside Group, U
(c) OrionX
Hybrid Algorithms
Quantum Approximate Optimization Algorithms
Hybrid Quantum Machine learning algorithms
Variational Quantum Algorithms
Quantum Approximate Optimization Algorithm
QAOA algorithms are used to solve combinatorial optimisation problems.
e.g. Max-cut problems
Max-cut problem: division of graphs into subsets such that the weights given to the
edges are maximised.
Unsolvable using classical algorithms
Max-cut problems
Quantum Approximate Optimization Algorithm
Step 1:
Start with random values of (γ, β)
Calculate HC , the objective function given by
where an edge ⟨i, j⟩ contributes to weight wi,j
and Xβi = exp(−iβiσ x)
Step 2:
Final state is measured to obtain expectation
values with respect to the objective function HC
L. Zhau, et. al. (2019)
∫ ψ * Oψdψ
Recall ⟨O⟩ =
∫ ψ * ψdψ
Step 3:
This expectation values are fed into classical optimiser to find the best parameters (γ, β)
L. Zhau, et. al. (2019)
An example problem is to have five vertex- graph as shown above. To find the configuration of spins such that the
weights between mis-aligned spins (black edges) are maximised.
Hybrid Quantum Machine learning algorithms
(c) Google AI blog
Hybrid Quantum CNN
https://doi.org/10.3390/e24030394
Variational Quantum Algorithms
Ansatz function is optimised using
where
(c) Nature Reviews (2021)
Applications of VQA’s
(c) Nature Reviews (2021)
