AI 용어집
인공지능 완전 사전
Variational Quantum Algorithm
Hybrid quantum-classical algorithm using a parameterized quantum circuit whose parameters are iteratively optimized by a classical optimizer to minimize a cost function.
Parameterized Quantum Circuit
Quantum circuit containing rotation gates with adjustable parameters that enable exploration of the quantum state space during variational optimization.
Quantum Ansatz
Predefined structure of parameterized quantum circuit chosen to approximate the solution of a computational problem, balancing expressibility and trainability.
Barren Plateau
Phenomenon where the cost function gradient becomes exponentially small with circuit depth, making optimization impossible in variational quantum algorithms.
Quantum Expressibility
Ability of a variational ansatz to generate a distribution of quantum states effectively covering the Hilbert space relevant to the considered problem.
Entangling Capability
Measure of a quantum circuit's ability to create entanglement between qubits, a crucial factor for the computational advantage of variational algorithms.
Quantum Cost Function Landscape
Topology of the objective function in the quantum parameter space, characterized by the presence of barren plateaus, local minima, and multi-modal landscapes.
Measurement Optimization
Strategies to reduce the number of measurements required in variational algorithms, including observable decomposition and importance sampling.
Trainability Depth Limit
Maximum depth beyond which a variational circuit becomes untrainable due to barren plateaus and accumulated quantum noise.
Variational Quantum Generator
Generative component in variational quantum algorithms, using a parameterized circuit to produce complex quantum probability distributions.
Ansatz Architecture Search
Process of automatically identifying the optimal ansatz circuit structure for a given problem, balancing expressibility and training complexity.