Variational Quantum Algorithms | Dr. Bharadwaj Mummameni

https://www.youtube.com/watch?v=iTNTapAqhKQ

Summary

TL;DR — This presentation introduces Variational Quantum Algorithms (VQAs), starting with a recap of quantum computing fundamentals like qubits, superposition, and entanglement, visualized using the Bloch sphere. It then delves into quantum gates, distinguishing between fixed and tunable rotation gates, and explains how these form quantum circuits and parameterized circuits called ansätze.

Key points

• The lecture covers fundamental quantum concepts: qubits, superposition, entanglement, and their visualization on the Bloch sphere.

• Quantum gates, both fixed (like X, H, Z) and tunable (RX, RY, RZ), are explained as operations on qubits, with an emphasis on their reversibility and role in forming quantum circuits.

• An 'ansatz' is defined as a parameterized quantum circuit, where tunable gate parameters (like rotation angles) are the free variables, allowing exploration of a subspace of quantum states.

• The necessity of ansätze arises from the vastness of the Hilbert space; they restrict the search to a manageable subspace, balancing parameter count with search efficiency.

• Two main types of ansätze are discussed: hardware-efficient (generic, shallow circuits, noisy hardware friendly) and problem-inspired (tailored, deeper circuits, requires domain expertise).

• A trade-off exists between expressibility (the range of states an ansatz can reach) and trainability (how easy it is to optimize), with a 'sweet spot' sought for optimal performance.

Takeaway — Variational Quantum Algorithms leverage parameterized quantum circuits (ansätze) to explore specific subspaces of quantum states, balancing expressibility and trainability to find solutions in a hybrid quantum-classical loop.