Barren plateaus in quantum neural network training landscapes
Jarrod R. McClean, Sergio Boixo, Vadim Smelyanskiy, Ryan Babbush, Hartmut Neven
Many experimental proposals for noisy intermediate scale quantum devices involve training a parameterized quantum circuit with a classical optimization loop. Such hybrid quantum-classical algorithms are popular for applications in quantum simulation, optimization, and machine learning. Due to its simplicity and hardware efficiency, random circuits are often proposed as initial guesses for exploring the space of quantum states. We show that the exponential dimension of Hilbert space and the gradient estimation complexity make this choice unsuitable for hybrid quantum-classical algorithms run on more than a few qubits. Specifically, we show that for a wide class of reasonable parameterized quantum circuits, the probability that the gradient along any reasonable direction is non-zero to some fixed precision is exponentially small as a function of the number of qubits. We argue that this is related to the 2-design characteristic of random circuits, and that solutions to this problem must be studied.
this paper
works it cites
works citing it
node size = global citations · hover for the full title
Citation edge evidence
- Quantum supremacy using a programmable superconducting processor→Barren plateaus in quantum neural network training landscapesLegacy edge — source detail unavailable
- Variational quantum algorithms→Barren plateaus in quantum neural network training landscapesLegacy edge — source detail unavailable
- Quantum supremacy using a programmable superconducting processor→Barren plateaus in quantum neural network training landscapesLegacy edge — source detail unavailable
- Noisy intermediate-scale quantum algorithms→Barren plateaus in quantum neural network training landscapesLegacy edge — source detail unavailable
- Quantum computational chemistry→Barren plateaus in quantum neural network training landscapesLegacy edge — source detail unavailable
- Quantum Chemistry in the Age of Quantum Computing→Barren plateaus in quantum neural network training landscapesLegacy edge — source detail unavailable
- Barren plateaus in quantum neural network training landscapes→Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Sh…Legacy edge — source detail unavailable
- Barren plateaus in quantum neural network training landscapes→A Fast Learning Algorithm for Deep Belief NetsLegacy edge — source detail unavailable
- Barren plateaus in quantum neural network training landscapes→Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Sh…Legacy edge — source detail unavailable
- Barren plateaus in quantum neural network training landscapes→Understanding the difficulty of training deep feedforward neural networksLegacy edge — source detail unavailable
- Barren plateaus in quantum neural network training landscapes→Quantum Computing in the NISQ era and beyondLegacy edge — source detail unavailable
- Barren plateaus in quantum neural network training landscapes→A variational eigenvalue solver on a photonic quantum processorLegacy edge — source detail unavailable
- Barren plateaus in quantum neural network training landscapes→Quantum machine learningLegacy edge — source detail unavailable
- Barren plateaus in quantum neural network training landscapes→Hardware-efficient variational quantum eigensolver for small molecules and quantum magnetsLegacy edge — source detail unavailable
- Barren plateaus in quantum neural network training landscapes→The theory of variational hybrid quantum-classical algorithmsLegacy edge — source detail unavailable
- Barren plateaus in quantum neural network training landscapes→A Quantum Approximate Optimization AlgorithmLegacy edge — source detail unavailable
- Understanding the difficulty of training deep feedforward neural networks→A Fast Learning Algorithm for Deep Belief NetsLegacy edge — source detail unavailable
- Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Sh…→Understanding the difficulty of training deep feedforward neural networksLegacy edge — source detail unavailable
- The theory of variational hybrid quantum-classical algorithms→A variational eigenvalue solver on a photonic quantum processorLegacy edge — source detail unavailable
- Quantum machine learning→A Quantum Approximate Optimization AlgorithmLegacy edge — source detail unavailable
- Hardware-efficient variational quantum eigensolver for small molecules and quantum magnets→A Quantum Approximate Optimization AlgorithmLegacy edge — source detail unavailable
- Hardware-efficient variational quantum eigensolver for small molecules and quantum magnets→A variational eigenvalue solver on a photonic quantum processorLegacy edge — source detail unavailable
- Hardware-efficient variational quantum eigensolver for small molecules and quantum magnets→The theory of variational hybrid quantum-classical algorithmsLegacy edge — source detail unavailable
- Quantum Computing in the NISQ era and beyond→A Quantum Approximate Optimization AlgorithmLegacy edge — source detail unavailable
- Quantum Computing in the NISQ era and beyond→A variational eigenvalue solver on a photonic quantum processorLegacy edge — source detail unavailable
- Quantum Computing in the NISQ era and beyond→The theory of variational hybrid quantum-classical algorithmsLegacy edge — source detail unavailable
- Quantum Computing in the NISQ era and beyond→Quantum machine learningLegacy edge — source detail unavailable
- Quantum computational chemistry→A Quantum Approximate Optimization AlgorithmLegacy edge — source detail unavailable
- Quantum computational chemistry→A variational eigenvalue solver on a photonic quantum processorLegacy edge — source detail unavailable
- Quantum computational chemistry→The theory of variational hybrid quantum-classical algorithmsLegacy edge — source detail unavailable
- Quantum computational chemistry→Hardware-efficient variational quantum eigensolver for small molecules and quantum magnetsLegacy edge — source detail unavailable
- Quantum computational chemistry→Quantum Computing in the NISQ era and beyondLegacy edge — source detail unavailable
- Quantum computational chemistry→Quantum Chemistry in the Age of Quantum ComputingLegacy edge — source detail unavailable
- Quantum computational chemistry→Quantum supremacy using a programmable superconducting processorLegacy edge — source detail unavailable
- Quantum computational chemistry→Quantum supremacy using a programmable superconducting processorLegacy edge — source detail unavailable
- Quantum Chemistry in the Age of Quantum Computing→A variational eigenvalue solver on a photonic quantum processorLegacy edge — source detail unavailable
- Quantum Chemistry in the Age of Quantum Computing→The theory of variational hybrid quantum-classical algorithmsLegacy edge — source detail unavailable
- Quantum Chemistry in the Age of Quantum Computing→Hardware-efficient variational quantum eigensolver for small molecules and quantum magnetsLegacy edge — source detail unavailable
- Quantum Chemistry in the Age of Quantum Computing→Quantum Computing in the NISQ era and beyondLegacy edge — source detail unavailable
- Quantum supremacy using a programmable superconducting processor→A variational eigenvalue solver on a photonic quantum processorLegacy edge — source detail unavailable
- Quantum supremacy using a programmable superconducting processor→Quantum Computing in the NISQ era and beyondLegacy edge — source detail unavailable
- Quantum supremacy using a programmable superconducting processor→A variational eigenvalue solver on a photonic quantum processorLegacy edge — source detail unavailable
- Quantum supremacy using a programmable superconducting processor→Quantum Computing in the NISQ era and beyondLegacy edge — source detail unavailable
- Variational quantum algorithms→A variational eigenvalue solver on a photonic quantum processorLegacy edge — source detail unavailable
- Variational quantum algorithms→The theory of variational hybrid quantum-classical algorithmsLegacy edge — source detail unavailable
- Variational quantum algorithms→Quantum machine learningLegacy edge — source detail unavailable
- Variational quantum algorithms→Hardware-efficient variational quantum eigensolver for small molecules and quantum magnetsLegacy edge — source detail unavailable
- Variational quantum algorithms→Quantum Computing in the NISQ era and beyondLegacy edge — source detail unavailable
- Variational quantum algorithms→Quantum supremacy using a programmable superconducting processorLegacy edge — source detail unavailable
- Noisy intermediate-scale quantum algorithms→A variational eigenvalue solver on a photonic quantum processorLegacy edge — source detail unavailable
- Noisy intermediate-scale quantum algorithms→The theory of variational hybrid quantum-classical algorithmsLegacy edge — source detail unavailable
- Noisy intermediate-scale quantum algorithms→Quantum machine learningLegacy edge — source detail unavailable
- Noisy intermediate-scale quantum algorithms→Hardware-efficient variational quantum eigensolver for small molecules and quantum magnetsLegacy edge — source detail unavailable
- Noisy intermediate-scale quantum algorithms→Quantum Computing in the NISQ era and beyondLegacy edge — source detail unavailable
- Noisy intermediate-scale quantum algorithms→Quantum computational chemistryLegacy edge — source detail unavailable
- Noisy intermediate-scale quantum algorithms→Quantum Chemistry in the Age of Quantum ComputingLegacy edge — source detail unavailable
- Noisy intermediate-scale quantum algorithms→Quantum supremacy using a programmable superconducting processorLegacy edge — source detail unavailable
What this paper cites, inside the corpus
What cites it, inside the corpus
Links
Download BibTeX · Download RIS · DOI · Open access copy · OpenAlex record
Topics
| Quantum Computing Algorithms and Architecture | Computer Science |
| Quantum Information and Cryptography | Computer Science |
| Stochastic Gradient Optimization Techniques | Computer Science |
Is this record sound?
complete
Nothing in this record contradicts itself and no field we check is missing.
- supports5 author record(s) attached.
- supports63 reference(s) recorded.
- neutralThe DOI carries no year to check against.
- supportsA title is present.
Provenance
sha256 86a6d6715ff540a0…