Toward Trainability of Quantum Neural Networks.

Quantum Neural Networks (QNNs) have been recently proposed as generalizations of classical neural networks to achieve the quantum speed-up. Despite the potential to outperform classical models, serious bottlenecks exist for training QNNs; namely, QNNs with random structures have poor trainability due to the vanishing gradient with rate exponential to the input qubit number. The vanishing gradient could seriously influence the applications of large-size QNNs. In this work, we provide a first viable solution with theoretical guarantees. Specifically, we prove that QNNs with tree tensor architectures have gradients that vanish polynomially with the qubit number. Moreover, our result holds irrespective of which encoding methods are employed. We numerically demonstrate QNNs with tree tensor structures for the application of binary classification. Simulations show faster convergent rates and better accuracy compared to QNNs with random structures.

[1]  Kohei Hayashi,et al.  Exploring Unexplored Tensor Network Decompositions for Convolutional Neural Networks , 2022, NeurIPS.

[2]  Simone Severini,et al.  Hierarchical quantum classifiers , 2018, npj Quantum Information.

[3]  Tobias J. Osborne,et al.  Training deep quantum neural networks , 2020, Nature Communications.

[4]  Arthur Pesah,et al.  Absence of Barren Plateaus in Quantum Convolutional Neural Networks , 2020, Physical Review X.

[5]  Nathan Killoran,et al.  PennyLane: Automatic differentiation of hybrid quantum-classical computations , 2018, ArXiv.

[6]  Kunal Sharma,et al.  Trainability of Dissipative Perceptron-Based Quantum Neural Networks , 2020, ArXiv.

[7]  Geoffrey E. Hinton,et al.  ImageNet classification with deep convolutional neural networks , 2012, Commun. ACM.

[8]  A. Harrow,et al.  Quantum algorithm for linear systems of equations. , 2008, Physical review letters.

[9]  Gang Su,et al.  Machine learning by unitary tensor network of hierarchical tree structure , 2017, New Journal of Physics.

[10]  Zhao Chen,et al.  GradNorm: Gradient Normalization for Adaptive Loss Balancing in Deep Multitask Networks , 2017, ICML.

[11]  Jingling Li,et al.  Final Report: Expressive Power of Parametrized Quantum Circuits , 2019 .

[12]  Guigang Zhang,et al.  Deep Learning , 2016, Int. J. Semantic Comput..

[13]  Marcello Benedetti,et al.  Parameterized quantum circuits as machine learning models , 2019, Quantum Science and Technology.

[14]  Robert Hecht-Nielsen,et al.  Theory of the backpropagation neural network , 1989, International 1989 Joint Conference on Neural Networks.

[15]  A. Harrow,et al.  Random Quantum Circuits are Approximate 2-designs , 2008, 0802.1919.

[16]  Ryan Babbush,et al.  Barren plateaus in quantum neural network training landscapes , 2018, Nature Communications.

[17]  Gavin E. Crooks,et al.  Gradients of parameterized quantum gates using the parameter-shift rule and gate decomposition , 2019, 1905.13311.

[18]  Edward Grant,et al.  An initialization strategy for addressing barren plateaus in parametrized quantum circuits , 2019, Quantum.

[19]  Travis S. Humble,et al.  Quantum supremacy using a programmable superconducting processor , 2019, Nature.

[20]  Hartmut Neven,et al.  Classification with Quantum Neural Networks on Near Term Processors , 2018, 1802.06002.

[21]  Dacheng Tao,et al.  On the learnability of quantum neural networks , 2020, 2007.12369.

[22]  Francesco Petruccione,et al.  Circuit-Based Quantum Random Access Memory for Classical Data , 2019, Scientific Reports.

[23]  Akira Sone,et al.  Cost-Function-Dependent Barren Plateaus in Shallow Quantum Neural Networks , 2020, ArXiv.

[24]  Frank Hutter,et al.  Neural Architecture Search: A Survey , 2018, J. Mach. Learn. Res..

[25]  Ievgeniia Oshurko.  Quantum Machine Learning , 2020, Quantum Computing.

[26]  K. Birgitta Whaley,et al.  Towards quantum machine learning with tensor networks , 2018, Quantum Science and Technology.

[27]  Leo Zhou,et al.  Quantum Approximate Optimization Algorithm: Performance, Mechanism, and Implementation on Near-Term Devices , 2018, Physical Review X.

[28]  M. Schuld,et al.  Circuit-centric quantum classifiers , 2018, Physical Review A.

[29]  Iordanis Kerenidis,et al.  Quantum Algorithms for Deep Convolutional Neural Networks , 2020, ICLR.

[30]  Masoud Mohseni,et al.  Layerwise learning for quantum neural networks , 2020, Quantum Machine Intelligence.

[31]  Robert König,et al.  Quantum advantage with shallow circuits , 2017, Science.

[32]  J. Gambetta,et al.  Hardware-efficient variational quantum eigensolver for small molecules and quantum magnets , 2017, Nature.

[33]  Ah Chung Tsoi,et al.  The Graph Neural Network Model , 2009, IEEE Transactions on Neural Networks.

[34]  Wojciech Zaremba,et al.  Recurrent Neural Network Regularization , 2014, ArXiv.