Training quantum neural networks using the Quantum Information Bottleneck method

• URL: http://arxiv.org/abs/2212.02600v2
• Date: Mon, 22 Jan 2024 17:08:53 GMT
• Title: Training quantum neural networks using the Quantum Information Bottleneck method
• Authors: Ahmet Burak Catli, Nathan Wiebe
• Abstract summary: We provide a concrete method for training a quantum neural network to maximize the relevant information about a property that is transmitted through the network. This is significant because it gives an operationally well founded quantity to optimize when training autoencoders for problems where the inputs and outputs are fully quantum.
• Score: 0.6768558752130311
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: We provide in this paper a concrete method for training a quantum neural network to maximize the relevant information about a property that is transmitted through the network. This is significant because it gives an operationally well founded quantity to optimize when training autoencoders for problems where the inputs and outputs are fully quantum. We provide a rigorous algorithm for computing the value of the quantum information bottleneck quantity within error $\epsilon$ that requires $O(\log^2(1/\epsilon) + 1/\delta^2)$ queries to a purification of the input density operator if its spectrum is supported on $\{0\}~\bigcup ~[\delta,1-\delta]$ for $\delta>0$ and the kernels of the relevant density matrices are disjoint. We further provide algorithms for estimating the derivatives of the QIB function, showing that quantum neural networks can be trained efficiently using the QIB quantity given that the number of gradient steps required is polynomial.

Related papers

• Quantum Resources for Pure Thermal Shadows [0.0]
  Calculating the properties of Gibbs states is an important task in Quantum Chemistry and Quantum Machine Learning. Previous work has proposed a quantum algorithm which predicts Gibbs state expectation values for $M$ observables from only $logM$ measurements. In this work, we perform resource analysis for the circuits used in this algorithm, finding that quantum signal processing contributes most significantly to gate count and depth as system size increases.
  arXiv  Detail & Related papers  (2024-09-09T16:40:21Z)
• SWAP-less Implementation of Quantum Algorithms [0.0]
  We present a formalism based on tracking the flow of parity quantum information to implement algorithms on devices with limited connectivity. We leverage the fact that entangling gates not only manipulate quantum states but can also be exploited to transport quantum information.
  arXiv  Detail & Related papers  (2024-08-20T14:51:00Z)
• Supervised binary classification of small-scale digits images with a trapped-ion quantum processor [56.089799129458875]
  We show that a quantum processor can correctly solve the basic classification task considered. With the increase of the capabilities quantum processors, they can become a useful tool for machine learning.
  arXiv  Detail & Related papers  (2024-06-17T18:20:51Z)
• Quantum circuit for multi-qubit Toffoli gate with optimal resource [6.727984016678534]
  We design new quantum circuits for the $n$-Toffoli gate and general multi-controlled unitary, which have only $O(log n)$-depth and $O(n)$-size. We demonstrate that without the assistance of ancillary qubit, any quantum circuit implementation of multi-qubit Toffoli gate must employ exponential precision gates.
  arXiv  Detail & Related papers  (2024-02-07T17:53:21Z)
• Training Multi-layer Neural Networks on Ising Machine [41.95720316032297]
  This paper proposes an Ising learning algorithm to train quantized neural network (QNN) As far as we know, this is the first algorithm to train multi-layer feedforward networks on Ising machines.
  arXiv  Detail & Related papers  (2023-11-06T04:09:15Z)
• Accelerating the training of single-layer binary neural networks using the HHL quantum algorithm [58.720142291102135]
  We show that useful information can be extracted from the quantum-mechanical implementation of Harrow-Hassidim-Lloyd (HHL) This paper shows, however, that useful information can be extracted from the quantum-mechanical implementation of HHL, and used to reduce the complexity of finding the solution on the classical side.
  arXiv  Detail & Related papers  (2022-10-23T11:58:05Z)
• Estimating Gibbs partition function with quantumClifford sampling [6.656454497798153]
  We develop a hybrid quantum-classical algorithm to estimate the partition function. Our algorithm requires only a shallow $mathcalO(1)$-depth quantum circuit. Shallow-depth quantum circuits are considered vitally important for currently available NISQ (Noisy Intermediate-Scale Quantum) devices.
  arXiv  Detail & Related papers  (2021-09-22T02:03:35Z)
• A quantum algorithm for training wide and deep classical neural networks [72.2614468437919]
  We show that conditions amenable to classical trainability via gradient descent coincide with those necessary for efficiently solving quantum linear systems. We numerically demonstrate that the MNIST image dataset satisfies such conditions. We provide empirical evidence for $O(log n)$ training of a convolutional neural network with pooling.
  arXiv  Detail & Related papers  (2021-07-19T23:41:03Z)
• Quantum Annealing Formulation for Binary Neural Networks [40.99969857118534]
  In this work, we explore binary neural networks, which are lightweight yet powerful models typically intended for resource constrained devices. We devise a quadratic unconstrained binary optimization formulation for the training problem. While the problem is intractable, i.e., the cost to estimate the binary weights scales exponentially with network size, we show how the problem can be optimized directly on a quantum annealer.
  arXiv  Detail & Related papers  (2021-07-05T03:20:54Z)
• A Co-Design Framework of Neural Networks and Quantum Circuits Towards Quantum Advantage [37.837850621536475]
  In this article, we present the co-design framework, namely QuantumFlow, to provide such a missing link. QuantumFlow consists of novel quantum-friendly neural networks (QF-Nets), a mapping tool (QF-Map) to generate the quantum circuit (QF-Circ) for QF-Nets, and an execution engine (QF-FB) Evaluation results show that QF-pNet and QF-hNet can achieve 97.10% and 98.27% accuracy, respectively.
  arXiv  Detail & Related papers  (2020-06-26T06:25:03Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.