Advances in Quantum Deep Learning: An Overview

Garg, Siddhant; Ramakrishnan, Goutham

doi:10.48550/arxiv.2005.04316

Cited by 16 publications

(15 citation statements)

References 69 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A number of quantum classifiers have been implemented recently, as reported in [50], and their performance is frequently compared to that of their classical counterparts. There have also been advancements in the field of quantum deep learning [51], with recent work by [52] implementing a quantum generative adversarial network on a superconducting quantum processor for learning and the generation of real-world handwritten digital images, and [53] implementing quantum convolutional neural networks (QCNN) for different applications; among them is the use of QCNN to create a quantum error correction method that is optimized for a particular error model. It is vital to highlight the combinatory "hybrid quantum-classical algorithms", which include recent advances, such as those seen in [54], where hybrid quantum variational autoencoder was applied to a representation learning task as well as the work of [55] implementing hybrid quantum-classical convolutional neural network on a Tetris dataset for classification.…”

Section: Emerging Quantum Machine Learning Technologymentioning

confidence: 99%

Artificial Intelligence Computing at the Quantum Level

Ayoade

Rivas

Orduz

2022

Data

View full text Add to dashboard Cite

The extraordinary advance in quantum computation leads us to believe that, in the not-too-distant future, quantum systems will surpass classical systems. Moreover, the field’s rapid growth has resulted in the development of many critical tools, including programmable machines (quantum computers) that execute quantum algorithms and the burgeoning field of quantum machine learning, which investigates the possibility of faster computation than traditional machine learning. In this paper, we provide a thorough examination of quantum computing from the perspective of a physicist. The purpose is to give laypeople and scientists a broad but in-depth understanding of the area. We also recommend charts that summarize the field’s diversions to put the whole field into context.

show abstract

Section: Emerging Quantum Machine Learning Technologymentioning

confidence: 99%

Artificial Intelligence Computing at the Quantum Level

Ayoade

Rivas

Orduz

2022

Data

View full text Add to dashboard Cite

show abstract

“…-92 -In the former case, we refer to designing qualitatively novel kinds of Quantum Artificial Neurons and Quantum Neural Networks, specially in the context of Quantum Machine Learning [112][113][114][115][116][117][118][119][120][121][122][123]. In the much-broader latter category, we envision applications in the interacting quantum systems which, without underlying neural network structures, feature various forms and even hierarchical levels of intelligent behaviour.…”

Section: One-qubit Purely-qmm-ues: Selected General Highlightsmentioning

confidence: 99%

Unitary Evolutions From Interacting Quantum Memories: Closed Quantum Systems Directing Themselves Using Their State Histories

Tavanfar¹,

Aliasghar²,

Pezzutto³

2022

Preprint

View full text Add to dashboard Cite

We propose, formulate and anlyze novel closed quantum systems whose unitary time evolutions, and internal interactions, emerge out of their linking quantum memories. In a closed system of the kind, the unitary evolution operator is updated, moment by moment, using the quantum state history as a compositional resource. The 'Quantum Memory Made' Hamiltonians (QMM-Hs) generating these unitary evolutions are Hermitian operators made of arbitrarily-chosen past-until-present density operators of the closed system, or its arbitrary subsystems. The time evolution is described by novel nonlocal and nonlinear von Neumann and Schrödinger equations. We establish that nontrivial Purely-QMM unitary evolutions are 'Robustly Non-Markovian', in the sense that the largest temporal distance between the compositional quantum memories admit finite lower bounds set by their interaction couplings. After general formulation and considerations, we take on the sufficiently-complex task of classifying the phases of one-qubit pure-state evolutions generated by specific choices of QMM-Hs, combining analytical methods with extensive numerical simulations, and using QMM two-point functions as natural signature probes. Analyzing Hamiltonians which are purely QMM, and those combined with the conventional Hamiltonians, we obtain and characterize specific families of analytical solutions, and then classify generic numerical solutions. We establish that QMM phase diagrams are outstandingly rich, containing wide classes of novel unitary evolutions with physically remarkable behaviours. Moreover, we show that QMM interactions give rise to novel purely internal dynamical phase transitions. We suggest several independent fundamental and applied domains and disciplines where QMM-UEs can be utilized advantageously.

show abstract

“…Great efforts have been made to develop a quantum based learning algorithms. In [22], authors conduct a comparative study of classic DL architectures with various Quantum-based learning architecture from a different perspective. The first challenge researchers encountered is how to represent classical data (binary states) with quantum states.…”

Section: Related Workmentioning

confidence: 99%

QuClassi: A Hybrid Deep Neural Network Architecture based on Quantum State Fidelity

Stein¹,

Baheri²,

Chen³

et al. 2021

Preprint

View full text Add to dashboard Cite

In the past decade, remarkable progress has been achieved in deep learning related systems and applications. In the post Moore's Law era, however, the limit of semiconductor fabrication technology along with the increasing data size have slowed down the development of learning algorithms. In parallel, the fast development of quantum computing has pushed it to the new ear. Google illustrates quantum supremacy by completing a specific task (random sampling problem), in 200 seconds, which is impracticable for the largest classical computers. Due to the limitless potential, quantum based learning is an area of interest, in hopes that certain systems might offer a quantum speedup. In this work, we propose a novel architecture QuClassi, a quantum neural network for both binary and multi-class classification. Powered by a quantum differentiation function along with a hybrid quantum-classic design, QuClassi encodes the data with a reduced number of qubits and generates the quantum circuit, pushing it to the quantum platform for the best states, iteratively. We conduct intensive experiments on both the simulator and IBM-Q quantum platform. The evaluation results demonstrate that QuClassi is able to outperform the state-of-the-art quantumbased solutions, Tensorflow-Quantum and QuantumFlow by up to 53.75% and 203.00% for binary and multi-class classifications. When comparing to traditional deep neural networks, QuClassi achieves a comparable performance with 97.37% fewer parameters.

show abstract

Advances in Quantum Deep Learning: An Overview

Cited by 16 publications

References 69 publications

Artificial Intelligence Computing at the Quantum Level

Artificial Intelligence Computing at the Quantum Level

Unitary Evolutions From Interacting Quantum Memories: Closed Quantum Systems Directing Themselves Using Their State Histories

QuClassi: A Hybrid Deep Neural Network Architecture based on Quantum State Fidelity

Contact Info

Product

Resources

About