Drift Control of High-Dimensional RBM: A Computational Method Based on Neural Networks.

(Proceedings of the National Academy of Sciences, 2018, 8505-8510), we develop and illustrate a simulation-based computational method that relies heavily on deep neural network technology. ... A system manager chooses a drift vector θ(t) at each time t based on the history of Z, and the cost rate at time t depends on both Z(t) and θ(t). ... Similar to us, the authors build on Han et al. [17] to solve their (high-dimensional) drift rate control problem. ...

arXiv:2309.11651v3 fatcat:uzfb4pskbfd2ffcbdzafioicyi

Open Access Multiple Versions

Here, we propose a method of generating weighted MAX-2-SAT instances inspired by the frustrated-loop algorithm used by the quantum annealing community. ... For the random-loop algorithm, we provide a thorough theoretical and empirical analysis on its mathematical properties from the perspective of frustration, and observe empirically a double phase transition ... In terms of machine learning, this means that the frustrated loop algorithm can be applied to a variety of neural network structures. ...

dblp:journals/jmlr/PeiMV20 fatcat:pydc34gyyvdvtmx2nk3j75cdne

Szczepanski

ACKNOWLEDGEMENTS The authors would like to thank Alexander Vergara and his team for offering the benchmark sensor drift dataset, and Lei Zhang and his team for offering the drifted E-nose dataset. ... This method is acknowledged to be a good way of training a deep network. Based on that, we will assume a 3-layer DBN model to formulate the learning procedure. ... methods, which means that the deep neural network is extremely effective in extenuating the influence of sensor drift. ...

doi:10.1109/access.2020.3006729 fatcat:4th3ykr6d5gghcla5hmsvpwy4m

DOAJ

a high-dimensional and nonlinear initial feature set. ... In order to solve this top problem, Restricted Boltzmann Machine (RBM), with the advantage of feature learning and selection for initial feature set, has been stacked layer by layer to realize a high-dimensional ... As far as the connection form of neural network, RBM can be considered as a stochastic neural network based on a probability graph model. ...

doi:10.1109/access.2020.2978112 fatcat:sqmmcz4rbzeyvkhtta7f2ujtvq

DOAJ

For each aspect, we elaborate on the update modes, detection modes, and adaptation drift types of concept drift adaptation methods. ... At the same time, deep learning is one of the core technologies of the fourth industrial revolution that have become vital in decision making. ... • SOM-based concept drift adaptation methods SOM is often applied to create low-dimensional (usually two-dimensional) representations of high-dimensional datasets, while maintaining the topology of ...

doi:10.3390/app13116515 fatcat:qufluurctvbjlf6g2gelbvbphe

DOAJ

A non-destructive identification method was developed here based on dropout deep belief network in multi-spectral data of ancient ceramic. ... Some weight and bias values trained by RBM were used to initialize a back propagation (BP) neural network. ... Conflicts of Interest: The authors declare no conflict of interest. ...

doi:10.3390/s21041318 pmid:33673248 fatcat:fezjgqrer5dfbcpl5k6pki7aya

DOAJ

on classical computers. ... The QC-based fault diagnosis model uses a quantum computing assisted generative training process followed by discriminative training to address the shortcomings of classical algorithms. ... network architecture used in the repeating sub-network of the QC-based fault diagnosis model that produces a high level abstraction of the input data 3.1. ...

arXiv:2003.00264v2 fatcat:hearascdzrak5pgvrlhylqbz2a

Multiple Versions

In this paper, we propose a neural network learning algorithm for solving eigenvalue problems and boundary value problems (BVPs) for elliptic operators and initial BVPs (IBVPs) of quasi-linear parabolic ... The method is based on the Martingale property in the stochastic representation for the eigenvalue/BVP/IBVP problems and martingale principle for optimal stochastic controls. ... DeepMartNet -a Martingale based neural network First, we will propose a neural network for computing eigenvalue and eigenfunction for elliptic operator in high dimensions as arising from quantum mechanics ...

arXiv:2307.11942v3 fatcat:bwubiulcmjcgphwqagpdvi6kde

Open Access Multiple Versions

In this article, we suggest an in-depth learning approach for adapting financial fraud through the use of convolution neural networks (CNN). ... Deep learning (DL)arises from the idea of a multi-type representation of the human brain that incorporates basic characteristics at the low level or high-level abstractions.Financial fraud was a big issue ... In this paper we use a method of neural networking, the main contribution of this paper, to detect fraudulent transactions. Our method of detecting adaptive fraud has been tested quantitatively. ...

doi:10.46647/ijetms.2022.v06i01.005 fatcat:h77a47rfrvbptdq5bv7xnta4kq

Deep Belief Networks (DBNs) DBNs are typical deep neural networks, which are stacked by a number of restricted Boltzmann machines (RBMs). The RBM structure is shown in Figure 1 . ... The other drawback of the above dimensionality reduction methods is that they are based on a statistical model, which assumes the probability distribution of the process variables first, and then sets ... As a result, the complexity of FDSR is O (3K 2 + L)MN 2 . Case Study The FDSR method based on a deep neural network was applied to the TE process. ...

doi:10.3390/app8050795 fatcat:wo7vxkumkrb2jezjfg5e4shdbe

DOAJ

These results underline the relevance of probabilistic methods in constructing neural networks and demonstrate the power of quantum computers, even with limited hardware capabilities. ... We provide a robust defence to adversarial attacks on discriminative algorithms. ... Peter was a great leader and mentor, which had great impact not only in the authors' careers, but in that of so many others. Peter, you would have enjoyed so much the advances we are witnessing. ...

arXiv:2012.11619v1 fatcat:hcnmgznvsffdbnygdm6ydvkx3q

Multiple Versions

Trajectory tracking control based on waypoint behavior is a promising way for unmanned surface vehicle (USV) to achieve autonomous navigation. ... Finally, two DNNs are connected in parallel with the control loop of USV to provide predictive supervision and auxiliary decision making for traditional control methods. ... Figure 9 summarizes the proposed method. e Construction of GB-RBM. e RBM is a neural network based on energy. ...

doi:10.1155/2021/8926738 fatcat:makt3imod5eepd35xzka5jfopq

DOAJ Szczepanski

Spiking neural networks provide a theoretically grounded means to test computational hypotheses on neurally plausible algorithms of reinforcement learning through numerical simulation. ... In this work, we use a spiking neural network model to approximate the free energy of a restricted Boltzmann machine and apply it to the solution of PORL problems with high-dimensional observations. ... Acknowledgments A part of this study is the result of "Bioinformatics for Brain Sciences" carried out under the Strategic Research Program for Brain Sciences by the Ministry of Education, Culture, Sports ...

doi:10.1371/journal.pone.0115620 pmid:25734662 pmcid:PMC4347982 fatcat:evcfufkhh5f2jjzywvqeu4duki

DOAJ

The deep belief neural network is designed using a restricted Boltzmann machine (RBM) so that the layer of RBM communicates with the previous and subsequent layers. ... The result is evaluated using a Cooja simulator based on the cloud environment. The ... Deep learning is the extension of a standard neural network with stacked hidden layers. Deep learning can handle very large, high-dimensional data with billions of parameters. ...

doi:10.32604/csse.2022.019940 fatcat:isxxxismfzd53hfsobsttqnb3i

In general, the proposed gas neural network, named GasNet, consists of: six convolutional blocks, each block consist of six layers; a pooling layer; and a fully-connected layer. ... Inspired by the great success of DCNN in the field of computer vision, we designed a DCNN with up to 38 layers. ... Author Contributions: The work presented in this paper is a collaborative development by all of the authors. ...

doi:10.3390/s18010157 pmid:29316723 pmcid:PMC5795481 fatcat:jxdeyladcrhldkyuwfwmc7tqqi

DOAJ

Drift Control of High-Dimensional RBM: A Computational Method Based on Neural Networks [article]

Preserved Fulltext

Other Versions

Generating Weighted MAX-2-SAT Instances with Frustrated Loops: an RBM Case Study

Preserved Fulltext

A Drift-Compensating Novel Deep Belief Classification Network to Improve Gas Recognition of Electronic Noses

Preserved Fulltext

A Fault Diagnosis Methodology based on Non-stationary Monitoring Signals by Extracting Features with Unknown Probability Distribution

Preserved Fulltext

Concept Drift Adaptation Methods under the Deep Learning Framework: A Literature Review

Preserved Fulltext

Dropout Deep Belief Network Based Chinese Ancient Ceramic Non-Destructive Identification

Preserved Fulltext

Quantum Computing Assisted Deep Learning for Fault Detection and Diagnosis in Industrial Process Systems [article]

Preserved Fulltext

Other Versions

DeepMartNet – A Martingale based Deep Neural Network Learning Algorithm for Eigenvalue/BVP Problems and Optimal Stochastic Controls [article]

Preserved Fulltext

Other Versions

Convolutional Neural Network Approach for Mobile Banking Fraudulent Transaction to Detect Financial Frauds

Preserved Fulltext

Fisher Discriminative Sparse Representation Based on DBN for Fault Diagnosis of Complex System

Preserved Fulltext

Defence against adversarial attacks using classical and quantum-enhanced Boltzmann machines [article]

Preserved Fulltext

Deep Learning-Based Trajectory Tracking Control forUnmanned Surface Vehicle

Preserved Fulltext

A Spiking Neural Network Model of Model-Free Reinforcement Learning with High-Dimensional Sensory Input and Perceptual Ambiguity

Preserved Fulltext

Insider Attack Detection Using Deep Belief Neural Network in Cloud Computing

Preserved Fulltext

Gas Classification Using Deep Convolutional Neural Networks

Preserved Fulltext