Variational Quantum Classifiers Through the Lens of the Hessian

Abstract

In quantum computing, the variational quantum algorithms (VQAs) are well suited for finding optimal combinations of things in specific applications ranging from chemistry all the way to finance. The training of VQAs with gradient descent optimization algorithm has shown a good convergence. At an early stage, the simulation of variational quantum circuits on noisy intermediate-scale quantum (NISQ) devices suffers from noisy outputs. Just like classical deep learning, it also suffers from vanishing gradient problems. It is a realistic goal to study the topology of loss landscape, to visualize the curvature information and trainability of these circuits in the existence of vanishing gradients. In this paper, we calculate the Hessian and visualize the loss landscape of variational quantum classifiers at different points in parameter space. The curvature information of variational quantum classifiers (VQC) is interpreted and the loss function's convergence is shown. It helps us better understand the behavior of variational quantum circuits to tackle optimization problems efficiently. We investigated the variational quantum classifiers via Hessian on quantum computers, starting with a simple 4-bit parity problem to gain insight into the practical behavior of Hessian, then thoroughly analyzed the behavior of Hessian's eigenvalues on training the variational quantum classifier for the Diabetes dataset. Finally, we show how the adaptive Hessian learning rate can influence the convergence while training the variational circuits.

arXiv comment13 pages, 9 figures

References in corpus (32)

• PyTorch: An Imperative Style, High-Performance Deep Learning Library
• Quantum Computing in the NISQ era and beyond
• A variational eigenvalue solver on a quantum processor
• Barren plateaus in quantum neural network training landscapes
• A Quantum Approximate Optimization Algorithm
• Quantum Circuit Learning
• Quantum Convolutional Neural Networks
• Cost Function Dependent Barren Plateaus in Shallow Parametrized Quantum Circuits
• Evaluating analytic gradients on quantum hardware
• The power of quantum neural networks
• The quest for a Quantum Neural Network
• Quantum autoencoders for efficient compression of quantum data
• Quantum Natural Gradient
• Variational Quantum Computation of Excited States
• An initialization strategy for addressing barren plateaus in parametrized quantum circuits
• Absence of Barren Plateaus in Quantum Convolutional Neural Networks
• Quantum Computation of Electronic Transitions using a Variational Quantum Eigensolver
• Subspace-search variational quantum eigensolver for excited states
• Trainability of Dissipative Perceptron-Based Quantum Neural Networks
• Empirical Analysis of the Hessian of Over-Parametrized Neural Networks
• On barren plateaus and cost function locality in variational quantum algorithms
• Experimental Realization of a Quantum Autoencoder: The Compression of Qutrits via Machine Learning
• Quantum neural network
• On the Universality of the Quantum Approximate Optimization Algorithm
• Characterizing the loss landscape of variational quantum circuits
• Universal Effectiveness of High-Depth Circuits in Variational Eigenproblems
• Abrupt Transitions in Variational Quantum Circuit Training
• Low depth mechanisms for quantum optimization
• Toward Trainability of Quantum Neural Networks
• The loss landscape of overparameterized neural networks
• Addressing hard classical problems with Adiabatically Assisted Variational Quantum Eigensolvers
• Implementing Entangled States on a Quantum Computer

Cited by in corpus (3)

• Quantum State Tomography using Quantum Machine Learning
• Framework for Learning and Control in the Classical and Quantum Domains
• ORQVIZ: Visualizing High-Dimensional Landscapes in Variational Quantum Algorithms