Trainability of Dissipative Perceptron-Based Quantum Neural Networks

Abstract

Several architectures have been proposed for quantum neural networks (QNNs), with the goal of efficiently performing machine learning tasks on quantum data. Rigorous scaling results are urgently needed for specific QNN constructions to understand which, if any, will be trainable at a large scale. Here, we analyze the gradient scaling (and hence the trainability) for a recently proposed architecture that we called dissipative QNNs (DQNNs), where the input qubits of each layer are discarded at the layer's output. We find that DQNNs can exhibit barren plateaus, i.e., gradients that vanish exponentially in the number of qubits. Moreover, we provide quantitative bounds on the scaling of the gradient for DQNNs under different conditions, such as different cost functions and circuit depths, and show that trainability is not always guaranteed.

5 + 21 pages, 3+2 figures, final version accepted for publication in Physical Review Letters

References in corpus (9)

• The power of quantum neural networks
• Noise-Induced Barren Plateaus in Variational Quantum Algorithms
• The quest for a Quantum Neural Network
• Absence of Barren Plateaus in Quantum Convolutional Neural Networks
• Large gradients via correlation in random parameterized quantum circuits
• Entanglement Devised Barren Plateau Mitigation
• Analyzing the barren plateau phenomenon in training quantum neural networks with the ZX-calculus
• Reformulation of the No-Free-Lunch Theorem for Entangled Data Sets
• Reinforcement Learning for Digital Quantum Simulation

Cited by in corpus (86)

• The Variational Quantum Eigensolver: a review of methods and best practices
• Challenges and Opportunities in Quantum Machine Learning
• Generalization in quantum machine learning from few training data
• 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
• Group-Invariant Quantum Machine Learning
• Equivalence of quantum barren plateaus to cost concentration and narrow gorges
• A Lie Algebraic Theory of Barren Plateaus for Deep Parameterized Quantum Circuits
• Variational Quantum State Eigensolver
• Higher Order Derivatives of Quantum Neural Networks with Barren Plateaus
• Quantum computing models for artificial neural networks
• Theory for Equivariant Quantum Neural Networks
• Quantum Methods for Neural Networks and Application to Medical Image Classification
• A comprehensive review of Quantum Machine Learning: from NISQ to Fault Tolerance
• Theoretical Guarantees for Permutation-Equivariant Quantum Neural Networks
• Layer VQE: A Variational Approach for Combinatorial Optimization on Noisy Quantum Computers
• Does provable absence of barren plateaus imply classical simulability?
• Understanding quantum machine learning also requires rethinking generalization
• Quantum Neural Network for Quantum Neural Computing
• On the practical usefulness of the Hardware Efficient Ansatz
• Variational Quantum Singular Value Decomposition
• Learning quantum data with the quantum Earth Mover's distance
• Subtleties in the trainability of quantum machine learning models
• Reformulation of the No-Free-Lunch Theorem for Entangled Data Sets
• Analytic theory for the dynamics of wide quantum neural networks
• Building spatial symmetries into parameterized quantum circuits for faster training
• The Presence and Absence of Barren Plateaus in Tensor-network Based Machine Learning
• Quantum Machine Learning for Finance
• Toward Trainability of Quantum Neural Networks
• Trainability barriers and opportunities in quantum generative modeling
• Quantum Phase Recognition via Quantum Kernel Methods
• Quantum Error Correction with Quantum Autoencoders
• Medical image classification via quantum neural networks
• Trainability Enhancement of Parameterized Quantum Circuits via Reduced-Domain Parameter Initialization
• Variational Quantum Classifiers Through the Lens of the Hessian
• Dynamical simulation via quantum machine learning with provable generalization
• Non-trivial symmetries in quantum landscapes and their resilience to quantum noise
• A semi-agnostic ansatz with variable structure for quantum machine learning
• Noisy intermediate-scale quantum algorithm for semidefinite programming
• Generalized Quantum Assisted Simulator
• Quantum Mixed State Compiling
• Quantum Capsule Networks
• On fundamental aspects of quantum extreme learning machines
• Variational quantum simulation: a case study for understanding warm starts
• Retrieving information from a black hole using quantum machine learning
• Learning shallow quantum circuits
• Tight and Efficient Gradient Bounds for Parameterized Quantum Circuits
• Training variational quantum algorithms with random gate activation
• The Quantum Path Kernel: a Generalized Quantum Neural Tangent Kernel for Deep Quantum Machine Learning
• Variational quantum algorithm for unconstrained black box binary optimization: Application to feature selection
• Efficient estimation of trainability for variational quantum circuits
• Inference-Based Quantum Sensing
• Isometric tensor network optimization for extensive Hamiltonians is free of barren plateaus
• The Unified Effect of Data Encoding, Ansatz Expressibility and Entanglement on the Trainability of HQNNs
• Effects of quantum resources on the statistical complexity of quantum circuits
• Quantum algorithms for scientific computing
• GSQAS: Graph Self-supervised Quantum Architecture Search
• Barren plateaus from learning scramblers with local cost functions
• Variational quantum state discriminator for supervised machine learning
• Mitigated barren plateaus in the time-nonlocal optimization of analog quantum-algorithm protocols
• Application of deep quantum neural networks to finance
• Using (1 + 1)D Quantum Cellular Automata for Exploring Collective Effects in Large Scale Quantum Neural Networks
• Resource frugal optimizer for quantum machine learning
• Barren plateaus are swamped with traps
• Absence of barren plateaus and scaling of gradients in the energy optimization of isometric tensor network states
• Quantum kernel learning Model constructed with small data
• Adaptive shot allocation for fast convergence in variational quantum algorithms
• Dissipative quantum many-body dynamics in (1+1)D quantum cellular automata and quantum neural networks
• Machine-Learning Insights into the Entanglement-trainability Correlation of Parameterized Quantum Circuits
• Efficient quantum-enhanced classical simulation for patches of quantum landscapes
• Quantum Extremal Learning
• Quadratic Quantum Speedup for Perceptron Training
• Non-linear classification capability of quantum neural networks due to emergent quantum metastability
• Learning Fourier series with parametrized quantum circuits
• Pitfalls when tackling the exponential concentration of parameterized quantum models
• Quantum Generative Training Using Rényi Divergences
• Quantum Hopfield Model with Dilute Memories
• The Accuracy vs. Sampling Overhead Trade-off in Quantum Error Mitigation Using Monte Carlo-Based Channel Inversion
• A general approach to quantum integration of cross sections in high-energy physics
• Mitigating Noise-Induced Gradient Vanishing in Variational Quantum Algorithm Training
• Adaptive Quantum Generative Training using an Unbounded Loss Function
• Encoding strongly-correlated many-boson wavefunctions on a photonic quantum computer: application to the attractive Bose-Hubbard model
• On quantum neural networks
• Batched Line Search Strategy for Navigating through Barren Plateaus in Quantum Circuit Training