Equivalence of quantum barren plateaus to cost concentration and narrow gorges

Abstract

Optimizing parameterized quantum circuits (PQCs) is the leading approach to make use of near-term quantum computers. However, very little is known about the cost function landscape for PQCs, which hinders progress towards quantum-aware optimizers. In this work, we investigate the connection between three different landscape features that have been observed for PQCs: (1) exponentially vanishing gradients (called barren plateaus), (2) exponential cost concentration about the mean, and (3) the exponential narrowness of minina (called narrow gorges). We analytically prove that these three phenomena occur together, i.e., when one occurs then so do the other two. A key implication of this result is that one can numerically diagnose barren plateaus via cost differences rather than via the computationally more expensive gradients. More broadly, our work shows that quantum mechanics rules out certain cost landscapes (which otherwise would be mathematically possible), and hence our results are interesting from a quantum foundations perspective.

arXiv comment7+5 pages, 2+1 figures, updated to published version

References in corpus (60)

• Quantum Computing in the NISQ era and beyond
• Quantum Machine Learning
• A variational eigenvalue solver on a quantum processor
• Variational Quantum Algorithms
• Barren plateaus in quantum neural network training landscapes
• The theory of variational hybrid quantum-classical algorithms
• A Quantum Approximate Optimization Algorithm
• Noisy intermediate-scale quantum (NISQ) algorithms
• 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
• Noise-Induced Barren Plateaus in Variational Quantum Algorithms
• Efficient Learning for Deep Quantum Neural Networks
• Exact and Approximate Unitary 2-Designs: Constructions and Applications
• The quest for a Quantum Neural Network
• Connecting ansatz expressibility to gradient magnitudes and barren plateaus
• Hybrid quantum-classical algorithms and quantum error mitigation
• Quantum Natural Gradient
• Chaos and complexity by design
• An initialization strategy for addressing barren plateaus in parametrized quantum circuits
• Evenly distributed unitaries: on the structure of unitary designs
• Quantum-assisted quantum compiling
• Quantum Data Hiding
• Absence of Barren Plateaus in Quantum Convolutional Neural Networks
• Layerwise learning for quantum neural networks
• Local random quantum circuits are approximate polynomial-designs
• Variational quantum simulation of general processes
• Variational Fast Forwarding for Quantum Simulation Beyond the Coherence Time
• Effect of barren plateaus on gradient-free optimization
• Stochastic gradient descent for hybrid quantum-classical optimization
• Variational Quantum State Diagonalization
• Noise Resilience of Variational Quantum Compiling
• Sequential minimal optimization for quantum-classical hybrid algorithms
• Trainability of Dissipative Perceptron-Based Quantum Neural Networks
• Large gradients via correlation in random parameterized quantum circuits
• Entanglement Devised Barren Plateau Mitigation
• Variational Quantum State Eigensolver
• Generalized Probability Theories: What determines the structure of quantum theory?
• Barren plateaus preclude learning scramblers
• Higher Order Derivatives of Quantum Neural Networks with Barren Plateaus
• An Adaptive Optimizer for Measurement-Frugal Variational Algorithms
• Low-depth gradient measurements can improve convergence in variational hybrid quantum-classical algorithms
• On barren plateaus and cost function locality in variational quantum algorithms
• Variational Quantum Fidelity Estimation
• Expressibility of the alternating layered ansatz for quantum computation
• Classical Optimizers for Noisy Intermediate-Scale Quantum Devices
• Quantum Assisted Simulator
• Learning to learn with quantum neural networks via classical neural networks
• A Universal Training Algorithm for Quantum Deep Learning
• Subtleties in the trainability of quantum machine learning models
• Variational Consistent Histories as a Hybrid Algorithm for Quantum Foundations
• Unitary designs from statistical mechanics in random quantum circuits
• The Presence and Absence of Barren Plateaus in Tensor-network Based Machine Learning
• Toward Trainability of Quantum Neural Networks
• Operator Sampling for Shot-frugal Optimization in Variational Algorithms
• A Jacobi Diagonalization and Anderson Acceleration Algorithm For Variational Quantum Algorithm Parameter Optimization
• Quantum Optimization for Training Quantum Neural Networks
• FLIP: A flexible initializer for arbitrarily-sized parametrized quantum circuits

Cited by in corpus (92)

• The Variational Quantum Eigensolver: a review of methods and best practices
• Challenges and Opportunities in Quantum Machine Learning
• Effect of barren plateaus on gradient-free optimization
• Diagnosing Barren Plateaus with Tools from Quantum Optimal Control
• Barren Plateaus in Variational Quantum Computing
• Quantum Computing for High-Energy Physics: State of the Art and Challenges. Summary of the QC4HEP Working Group
• Theory of overparametrization in quantum neural networks
• A Lie Algebraic Theory of Barren Plateaus for Deep Parameterized Quantum Circuits
• Theory for Equivariant Quantum Neural Networks
• A comprehensive review of Quantum Machine Learning: from NISQ to Fault Tolerance
• Theoretical Guarantees for Permutation-Equivariant Quantum Neural Networks
• Biology and medicine in the landscape of quantum advantages
• ADAPT-VQE is insensitive to rough parameter landscapes and barren plateaus
• Does provable absence of barren plateaus imply classical simulability?
• On the practical usefulness of the Hardware Efficient Ansatz
• Towards adiabatic quantum computing using compressed quantum circuits
• Subtleties in the trainability of quantum machine learning models
• The Adjoint Is All You Need: Characterizing Barren Plateaus in Quantum Ansätze
• Synergy Between Quantum Circuits and Tensor Networks: Short-cutting the Race to Practical Quantum Advantage
• The Presence and Absence of Barren Plateaus in Tensor-network Based Machine Learning
• Graph neural network initialisation of quantum approximate optimisation
• Towards large-scale quantum optimization solvers with few qubits
• ResQNets: A Residual Approach for Mitigating Barren Plateaus in Quantum Neural Networks
• Trainability barriers and opportunities in quantum generative modeling
• Mitigating Barren Plateaus with Transfer-learning-inspired Parameter Initializations
• Absence of barren plateaus in finite local-depth circuits with long-range entanglement
• An Expressive Ansatz for Low-Depth Quantum Approximate Optimisation
• Quantum Mixed State Compiling
• Entangled Datasets for Quantum Machine Learning
• On fundamental aspects of quantum extreme learning machines
• Entanglement entropy production in Quantum Neural Networks
• Quantum Capsule Networks
• Microwave signal processing using an analog quantum reservoir computer
• Variational quantum simulation: a case study for understanding warm starts
• Engineered dissipation to mitigate barren plateaus
• Optimal training of variational quantum algorithms without barren plateaus
• Tight and Efficient Gradient Bounds for Parameterized Quantum Circuits
• Efficient estimation of trainability for variational quantum circuits
• Neural network encoded variational quantum algorithms
• Quantum Convolutional Neural Networks are Effectively Classically Simulable
• Numerical evidence against advantage with quantum fidelity kernels on classical data
• Lie-algebraic classical simulations for quantum computing
• A General Approach to Dropout in Quantum Neural Networks
• Classical Splitting of Parametrized Quantum Circuits
• Isometric tensor network optimization for extensive Hamiltonians is free of barren plateaus
• Networked Quantum Services
• Quantum neural networks form Gaussian processes
• Quantum algorithms for scientific computing
• Connecting geometry and performance of two-qubit parameterized quantum circuits
• Quantum Energy Landscape and VQA Optimization
• Beyond unital noise in variational quantum algorithms: noise-induced barren plateaus and limit sets
• Mitigated barren plateaus in the time-nonlocal optimization of analog quantum-algorithm protocols
• Non-Unitary Quantum Machine Learning
• Constrained and Vanishing Expressivity of Quantum Fourier Models
• The battle of clean and dirty qubits in the era of partial error correction
• Large-scale simulations of Floquet physics on near-term quantum computers
• Quantum variational learning for entanglement witnessing
• F-Divergences and Cost Function Locality in Generative Modelling with Quantum Circuits
• Emergence of noise-induced barren plateaus in arbitrary layered noise models
• Characterization of variational quantum algorithms using free fermions
• Equivalence of cost concentration and gradient vanishing for quantum circuits: An elementary proof in the Riemannian formulation
• Barren plateaus are swamped with traps
• Absence of barren plateaus and scaling of gradients in the energy optimization of isometric tensor network states
• Quantum Chaos and Circuit Parameter Optimization
• Solving Fractional Differential Equations on a Quantum Computer: A Variational Approach
• Adaptive shot allocation for fast convergence in variational quantum algorithms
• Efficient Learning for Linear Properties of Bounded-Gate Quantum Circuits
• QSlack: A slack-variable approach for variational quantum semi-definite programming
• Efficient quantum-enhanced classical simulation for patches of quantum landscapes
• On the similarity of bandwidth-tuned quantum kernels and classical kernels
• Imaginary Hamiltonian variational ansatz for combinatorial optimization problems
• Pitfalls when tackling the exponential concentration of parameterized quantum models
• Dual-VQE: A quantum algorithm to lower bound the ground-state energy
• Learning Equivariant Maps with Variational Quantum Circuits
• Almost fault-tolerant quantum machine learning with drastic overhead reduction
• A Variational Quantum Algorithm For Approximating Convex Roofs
• Normalized Gradient Descent for Variational Quantum Algorithms
• Adiabatic quantum computing with parameterized quantum circuits
• Benchmarking Variational Quantum Algorithms for Combinatorial Optimization in Practice
• Approximate Quadratization of High-Order Hamiltonians for Combinatorial Quantum Optimization
• Efficient Online Quantum Circuit Learning with No Upfront Training
• A coherent approach to quantum-classical optimization
• The Dual Role of Low-Weight Pauli Propagation: A Flawed Simulator but a Powerful Initializer for Variational Quantum Algorithms
• Warm Start of Variational Quantum Algorithms for Quadratic Unconstrained Binary Optimization Problems
• ORQVIZ: Visualizing High-Dimensional Landscapes in Variational Quantum Algorithms
• Exploiting many-body localization for scalable variational quantum simulation
• Interpolation-based coordinate descent method for parameterized quantum circuits
• Equivalence between exponential concentration in quantum machine learning kernels and barren plateaus in variational algorithms
• Natural gradient and parameter estimation for quantum Boltzmann machines
• Fast suppression of classification error in variational quantum circuits
• Hamiltonian Expressibility for Ansatz Selection in Variational Quantum Algorithms
• Improving the trainability of VQE on NISQ computers for solving portfolio optimization using convex interpolation