Search

Efficiently solving combinatorial optimization problems (COPs) such as Max‐Cut is challenging because the resources required increase exponentially with the problem size. This study proposes a hardware‐friendly method for solving the Max‐Cut problem by implementing a spiking neural network (SNN)‐based Boltzmann machine (BM) in neuromorphic hardware systems. To implement the hardware‐oriented version of the spiking Boltzmann machine (sBM), the stochastic dynamics of leaky integrate‐and‐fire (LIF) neurons with random walk noise are analyzed, and an innovative algorithm based on overlapping time windows is proposed. The simulation results demonstrate the effective convergence and high accuracy of the proposed method for large‐scale Max‐Cut problems. The proposed method is validated through successful hardware implementation on a 6‐transistor/2‐resistor (6T2R) neuromorphic chip with phase change memory (PCM) synapses. In addition, as an expansion of the algorithm, several annealing techniques and bias split methods are proposed to improve convergence, along with circuit design ideas for efficient evaluation of sampling convergence using cell arrays and spiking systems. Overall, the results of the proposed methods demonstrate the potential of energy‐efficient and hardware‐implementable approaches using SNNs to solve COPs. To the best of the author's knowledge, this is the first study to solve the Max‐Cut problem using an SNN neuromorphic hardware chip. [ABSTRACT FROM AUTHOR]

The purpose of this paper is to describe a system of interconnected energy models. These models produce inferences that may enhance our understanding of classical and quantum mechanics. From the formulated models, an empirical test is derived to validate the proposed light projection hypothesis. [ABSTRACT FROM AUTHOR]

A mixed-integer linear programming (MILP) formulation is presented for parameter estimation of the Potts model. Two algorithms are developed; the first method estimates the parameters such that the set of ground states replicate the user-prescribed data set; the second method allows the user to prescribe the ground states multiplicity. In both instances, the optimization process ensures that the bandgap is maximized. Consequently, the model parameter efficiently describes the user data for a broad range of temperatures. This is useful in the development of energy-based graph models to be simulated on Quantum annealing hardware where the exact simulation temperature is unknown. Computationally, the memory requirement in this method grows exponentially with the graph size. Therefore, this method can only be practically applied to small graphs. Such applications include learning of small generative classifiers and spin-lattice model with energy described by Ising hamiltonian. Learning large data sets poses no extra cost to this method; however, applications involving the learning of high dimensional data are out of scope. [ABSTRACT FROM AUTHOR]

In this article we present a concrete design for a probabilistic (p-) computer based on a network of p-bits, robust classical entities fluctuating between -1 and +1, with probabilities that are controlled through an input constructed from the outputs of other p-bits. The architecture of this probabilistic computer is similar to a stochastic neural network with the p-bit playing the role of a binary stochastic neuron, but with one key difference: there is no sequencer used to enforce an ordering of p-bit updates, as is typically required. Instead, we explore sequencerless designs where all p-bits are allowed to flip autonomously and demonstrate that such designs can allow ultrafast operation unconstrained by available clock speeds without compromising the solution's fidelity. Based on experimental results from a hardware benchmark of the autonomous design and benchmarked device models, we project that a nanomagnetic implementation can scale to achieve petaflips per second with millions of neurons. A key contribution of this article is the focus on a hardware metric - flips per second - as a problem and substrate-independent figure-of-merit for an emerging class of hardware annealers known as Ising Machines. Much like the shrinking feature sizes of transistors that have continually driven Moore's Law, we believe that flips per second can be continually improved in later technology generations of a wide class of probabilistic, domain specific hardware.

In this article we present a concrete design for a probabilistic (p-) computer based on a network of p-bits, robust classical entities fluctuating between -1 and +1, with probabilities that are controlled through an input constructed from the outputs of other p-bits. The architecture of this probabilistic computer is similar to a stochastic neural network with the p-bit playing the role of a binary stochastic neuron, but with one key difference: there is no sequencer used to enforce an ordering of p-bit updates, as is typically required. Instead, we explore sequencerless designs where all p-bits are allowed to flip autonomously and demonstrate that such designs can allow ultrafast operation unconstrained by available clock speeds without compromising the solution's fidelity. Based on experimental results from a hardware benchmark of the autonomous design and benchmarked device models, we project that a nanomagnetic implementation can scale to achieve petaflips per second with millions of neurons. A key contribution of this article is the focus on a hardware metric - flips per second - as a problem and substrate-independent figure-of-merit for an emerging class of hardware annealers known as Ising Machines. Much like the shrinking feature sizes of transistors that have continually driven Moore's Law, we believe that flips per second can be continually improved in later technology generations of a wide class of probabilistic, domain specific hardware.

Boltzmann machines (BMs) are graphical models with interconnected binary units, employed for the unsupervised modeling of data distributions. When trained on real data, BMs show the tendency to behave like critical systems, displaying a high susceptibility of the model under a small rescaling of the inferred parameters. This behavior is not convenient for the purpose of generating data, because it slows down the sampling process, and induces the model to overfit the training-data. In this study, we introduce a regularization method for BMs to improve the robustness of the model under rescaling of the parameters. The new technique shares formal similarities with the unlearning algorithm, an iterative procedure used to improve memory associativity in Hopfield-like neural networks. We test our unlearning regularization on synthetic data generated by two simple models, the Curie–Weiss ferromagnetic model and the Sherrington–Kirkpatrick spin glass model. We show that it outperforms L _p -norm schemes and discuss the role of parameter initialization. Eventually, the method is applied to learn the activity of real neuronal cells, confirming its efficacy at shifting the inferred model away from criticality and coming out as a powerful candidate for actual scientific implementations.

In this study, we explore the similarities and differences between variational Monte Carlo techniques that employ conventional and artificial neural network representations of the ground-state wave function for fermionic systems. Our primary focus is on shallow neural network architectures, specifically the restricted Boltzmann machine, and we examine unsupervised learning algorithms that are appropriate for modeling complex many-body correlations. We assess the advantages and drawbacks of conventional and neural network wave functions by applying them to a range of circular quantum dot systems. Our findings, which include results for systems containing up to 90 electrons, emphasize the efficient implementation of these methods on both homogeneous and heterogeneous high-performance computing facilities.

This paper presents a collaborative neurodynamic approach to Boolean matrix factorization. Based on a binary optimization formulation to minimize the Hamming distance between a given data matrix and its low-rank reconstruction, the proposed approach employs a population of Boltzmann machines operating concurrently for scatter search of factorization solutions. In addition, a particle swarm optimization rule is used to re-initialize the neuronal states of Boltzmann machines upon their local convergence to escape from local minima toward global solutions. Experimental results demonstrate the superior convergence and performance of the proposed approach against six baseline methods on ten benchmark datasets. [ABSTRACT FROM AUTHOR]

The world is advancing to a new era where a new concept is emerging that deals with "wirelessness". As we know, renewable energy is the future, and this research studied the integration of both fields that results in a futuristic, powerful, and advanced model of wireless body area networks. Every new emerging technology does have some cons; in this case the issue would be the usage of excess energy by the sensors of the model. Our research is focused on solving this excessive usage of energy to promote the optimization of energy. This research work is aimed to design a power-saving protocol (PSP) for wireless body area networks (WBANs) in electronic health monitoring (EHM). Our proposed power-saving protocol (PSP) supports the early detection of suspicious signs or sporadic elder movements. The protocol focuses on solving the excessive energy consumption by the body attached to IoT devices to maximize the power efficiency (EE) of WBAN. In a WSNs network, the number of sensor nodes (SNs) interact with an aggregator and are equipped with energy harvesting capabilities. The energy optimization for the wireless sensor networks is a vital step and the methodology is completely based on renewable energy resources. Our proposed power-saving protocol is based on AI and DNN architectures with a hidden Markov model to obtain the top and bottom limits of the SN sources and a less computationally challenging suboptimal elucidation. The research also addressed many critical technical problems, such as sensor node hardware configuration and energy conservation. The study performed the simulation using the OMNET++ environment and represent through results the source rate to power critical SNs improves WBAN's scheme performance in terms of power efficiency of Sporadic Elder Movements (SEM) during various daily operations. [ABSTRACT FROM AUTHOR]

Abstract The phenomenon of two or more genes affecting the expression of each other in various ways in the development of a single character of an organism is known as gene interaction. Gene interaction not only applies to normal human traits but to the diseased samples as well. Thus, an analysis of gene interaction could help us to differentiate between the normal and the diseased samples or between the two/more phases any diseased samples. At the first stage of this work we have used restricted Boltzmann machine model to find such significant interactions present in normal and/or cancer samples of every gene pairs of 20 genes of colorectal cancer data set (GDS4382) along with the weight/degree of those interactions. Later on, we are looking for those interactions present in adenoma and/or carcinoma samples of the same 20 genes of colorectal cancer data set (GDS1777). The weight/degree of those interactions represents how strong/weak an interaction is. At the end we will create a gene regulatory network with the help of those interactions, where the regulatory genes are identified by using Naïve Bayes Classifier. Experimental results are validated biologically by comparing the interactions with NCBI databases.

In the past, combinatorial structures have been used only to tune parameters of neural networks. In this work, we employ neural networks in the form of Boltzmann machines and Hopfield networks for the construction of a specific class of combinatorial designs, namely covering arrays (CAs). In past works, these neural networks were successfully used to solve set cover instances. For the construction of CAs, we consider the corresponding set cover instances and use neural networks to solve such instances. We adapt existing algorithms for solving general set cover instances, which are based on Boltzmann machines and Hopfield networks and apply them for CA construction. Furthermore, for the algorithm based on Boltzmann machines, we consider newly designed versions, where we deploy structural changes of the underlying Boltzmann machine, adding a feedback loop. Additionally, one variant of this algorithm employs learning techniques based on neural networks to adjust the various connections encountered in the graph representing the considered set cover instances. Culminating in a comprehensive experimental evaluation, our work presents the first study of applications of neural networks in the field of covering array generation and related discrete structures and may act as a guideline for future investigations. [ABSTRACT FROM AUTHOR]

Tensor networks have found a wide use in a variety of applications in physics and computer science, recently leading to both theoretical insights as well as practical algorithms in machine learning. In this work we explore the connection between tensor networks and probabilistic graphical models, and show that it motivates the definition of generalized tensor networks where information from a tensor can be copied and reused in other parts of the network. We discuss the relationship between generalized tensor network architectures used in quantum physics, such as string-bond states, and architectures commonly used in machine learning. We provide an algorithm to train these networks in a supervised-learning context and show that they overcome the limitations of regular tensor networks in higher dimensions, while keeping the computation efficient. A method to combine neural networks and tensor networks as part of a common deep learning architecture is also introduced. We benchmark our algorithm for several generalized tensor network architectures on the task of classifying images and sounds, and show that they outperform previously introduced tensor-network algorithms. The models we consider also have a natural implementation on a quantum computer and may guide the development of near-term quantum machine learning architectures.

With an exponential increase in the amount of data collected per day, the fields of artificial intelligence and machine learning continue to progress at a rapid pace with respect to algorithms, models, applications, and hardware. In particular, deep neural networks have revolutionized these fields by providing unprecedented human-like performance in solving many real-world problems such as image or speech recognition. There is also significant research aimed at unraveling the principles of computation in large biological neural networks and, in particular, biologically plausible spiking neural networks. This article presents an overview of the brain-inspired computing models starting with the development of the perceptron and multi-layer perceptron followed by convolutional neural networks (CNNs) and recurrent neural networks (RNNs). This article also briefly reviews other neural network models such as Hopfield neural networks and Boltzmann machines. Other models such as spiking neural networks (SNNs) and hyperdimensional computing are then briefly reviewed. Recent advances in these neural networks and graph related neural networks are then described.

In this article we present a concrete design for a probabilistic (p-) computer based on a network of p-bits, robust classical entities fluctuating between -1 and +1, with probabilities that are controlled through an input constructed from the outputs of other p-bits. The architecture of this probabilistic computer is similar to a stochastic neural network with the p-bit playing the role of a binary stochastic neuron, but with one key difference: there is no sequencer used to enforce an ordering of p-bit updates, as is typically required. Instead, we explore sequencerless designs where all p-bits are allowed to flip autonomously and demonstrate that such designs can allow ultrafast operation unconstrained by available clock speeds without compromising the solution's fidelity. Based on experimental results from a hardware benchmark of the autonomous design and benchmarked device models, we project that a nanomagnetic implementation can scale to achieve petaflips per second with millions of neurons. A key contribution of this article is the focus on a hardware metric - flips per second - as a problem and substrate-independent figure-of-merit for an emerging class of hardware annealers known as Ising Machines. Much like the shrinking feature sizes of transistors that have continually driven Moore's Law, we believe that flips per second can be continually improved in later technology generations of a wide class of probabilistic, domain specific hardware.

Sensor‐based human activity classification requires time and frequency domain feature extraction techniques. The set of choice in time and frequency domain features may have a significant impact on the overall classification accuracy. Another problem is to train deep learning models with sufficient dataset. The use of generative models eliminates the requirement of choosing certain features of the signal. As a generative model, restricted Boltzmann machine (RBM) is an energy‐based probabilistic graphical model which factorises the probability distribution of a random variable over a binary probability distribution. Conditional restricted Boltzmann machines (CRBMs) is an extension to RBM, which can capture temporal information in time‐series signals and can be deployed as a generative model in classification. In this study, the authors show how CRBMs can be trained to learn signal features. They present four generative model training results, RBM, CRBM, generative adversarial network, Wasserstein generative adversarial network – gradient penalty and compare the models' performances with a performance criterion. They show that the CRBM model can generate signals closest to true signals with a significantly higher success rate as compared to other presented generative models. They present a statistical analysis of the findings and show that the findings significantly hold. [ABSTRACT FROM AUTHOR]

Fisher kernels derived from stochastic probabilistic models such as restricted and deep Boltzmann machines have shown competitive visual classification results in comparison to widely popular deep discriminative models. This genre of Fisher kernels bridges the gap between shallow and deep learning paradigm by inducing the characteristics of deep architecture into Fisher kernel, further deployed for classification in discriminative classifiers. Despite their success, the memory and computational costs of Fisher vectors do not make them amenable for large‐scale visual retrieval and classification tasks. This study introduces a novel feature selection technique inspired from the functional characteristics of neural architectures for learning discriminative feature representations to boost the performance of Fisher kernels against deep discriminative models. The proposed technique condenses the large dimensional Fisher features for kernel learning and shows improvement in its classification performance and storage cost on leading benchmark data sets. A comparison of the proposed method with other state‐of‐the‐art feature selection techniques is made to demonstrate its performance supremacy as well as time complexity required to learn in reduced Fisher space. [ABSTRACT FROM AUTHOR]

In this paper we study the properties of the quenched pressure of a multi-layer spin-glass model (a deep Boltzmann Machine in artificial intelligence jargon) whose pairwise interactions are allowed between spins lying in adjacent layers and not inside the same layer nor among layers at distance larger than one. We prove a theorem that bounds the quenched pressure of such a K-layer machine in terms of K Sherrington–Kirkpatrick spin glasses and use it to investigate its annealed region. The replica-symmetric approximation of the quenched pressure is identified and its relation to the annealed one is considered. The paper also presents some observation on the model's architectural structure related to machine learning. Since escaping the annealed region is mandatory for a meaningful training, by squeezing such region we obtain thermodynamical constraints on the form factors. Remarkably, its optimal escape is achieved by requiring the last layer to scale sub-linearly in the network size. [ABSTRACT FROM AUTHOR]

A general Boltzmann machine with continuous visible and discrete integer valued hidden states is introduced. Under mild assumptions about the connection matrices, the probability density function of the visible units can be solved for analytically, yielding a novel parametric density function involving a ratio of Riemann-Theta functions. The conditional expectation of a hidden state for given visible states can also be calculated analytically, yielding a derivative of the logarithmic Riemann-Theta function. The conditional expectation can be used as activation function in a feedforward neural network, thereby increasing the modelling capacity of the network. Both the Boltzmann machine and the derived feedforward neural network can be successfully trained via standard gradient- and non-gradient-based optimization techniques. [ABSTRACT FROM AUTHOR]

Cable incipient fault is an intermittent arc fault, and may evolve into a permanent fault. Due to the short duration of the fault, the conventional overcurrent protection device cannot detect it. A cable incipient fault identification method is proposed in this study, using restricted Boltzmann machine (RBM) and stacked autoencoder (SAE). Firstly, disturbance current waveforms data is effectively compressed by RBM, which can improve analysis efficiency and obtain the shallow features of the data. Then, the compressed data is used as the input of SAE, and the optimal network parameters are obtained through layer‐by‐layer pre‐training and fine‐tuning. Finally, a well‐trained SAE network is used to learn deep features from the input data to identify cable incipient fault, and softmax outputs identification result. In addition, the performance of the proposed method is compared with other methods. The accuracy of the proposed method is 98.33/95.62% for simulated data/measured data, and is 1.66/1.09%, 3.33/1.76%, 17.31/28.48% and 40.17/46.1% higher than the accuracies of convolutional neural network, deep belief network, random forest and back propagation neural network, respectively. The results show that the proposed method has high identification accuracy and feasibility. [ABSTRACT FROM AUTHOR]

Flow-based generative models (flow models) are conceptually attractive due to tractability of the exact log-likelihood and the exact latent-variable inference. In order to generate sharper images and extend the Gaussian prior of Flow models to other discrete forms, we propose an adversarial non-volume preserving flow model with Boltzmann priors (ANVP) for modeling complex high-dimensional densities. In order to generate sharper images, an ANVP model introduces an adversarial regularizer into the loss function to penalize the condition that it places a high probability in regions where the training data distribution has a low density. Moreover, we show that the Gaussian prior can be extended to other forms such as the Boltzmann prior in the proposed ANVP model, and we use multi-scale transformations and Boltzmann priors to model the data distribution. The experiments show that proposed model is effective in image generation task. [ABSTRACT FROM AUTHOR]

In this study, an inverse synthetic aperture radar (ISAR) image resolution enhancement algorithm based on joint dictionary learning is proposed, by which two special sets of sparse signals called dictionaries are solved by exploiting numerous high-resolution (HR) and low-resolution (LR) ISAR images. Herein a new coupled dictionary learning algorithm based on restricted Boltzmann machine (RBM) is designed to learn a LR and a HR dictionary using LR and HR image patches. Since the echoes are equivalent to similar scattering-centre models when an object is illuminated by radar signals with same centre frequency and different bandwidth, respectively, it is reasonable to assume the object's LR ISAR image shares the same sparse representation coefficients with its HR ISAR image. When a LR ISAR image is represented sparsely with a LR dictionary, a HR ISAR images can be reconstructed based on a HR dictionary owing to the similar sparse representation coefficients. Experiment results with simulation data demonstrate the superior performance of the proposed method over other classical dictionary training algorithms.

In this research work, a deep learning algorithm is applied to the medical domain to deliver a better healthcare system. For this, a deep learning framework for classification the region of interest pattern of complex hyperspectral medical images is proposed. The performance of computer-aided diagnosis by verifying the region in hyperspectral image by pre and post-cancerous region classification is enhanced. For this a deep Boltzmann machine (DBM) architecture of the bipartite structure as an unsupervised generative model was developed. The performance of DBM is compared with deep convolutional neural network architecture. For implementation, a three-layer unsupervised network with a backpropagation structure is used. From the presented dataset, image patches are collected and classified into two classes, namely non-informative and discriminative classes as labelled classes. The spatial information is used for classification and spectral-spatial representation of class labels is formed. In the labelled classes, the accuracy, false-positive predictions, sensitivity are obtained for the proposed fully-connected network. By the proposed cognitive computation technique an accuracy of 95.5% with 93.5% sensitivity was obtained. From the obtained classification, accuracy and success rate DBM provide a better classification of complex images compared to traditional convolution network.

We propose an integrated photonic circuit that acts as an optical coherent Ising machine and simulates its performance on the basis of some example problems. In contrast to previous all-optical approaches, the proposed integrated Ising machine does not require an optical parametric oscillator and can, hence, operate at a single wavelength, reducing the overall design complexity. We present a symmetric nonlinear photonic device as the fundamental building block that can use self-phase modulation in two microring resonators to emulate a continuously tunable, symmetrically bistable Ising node. We derive and verify using numerical simulation the key properties of the single Ising node device and the full Ising machine circuit. We estimate the full Ising machine's tolerance to realistic fabrication errors on the basis of randomly sampled example problems, and we discuss which technologies are required to obtain large-scale systems. [ABSTRACT FROM AUTHOR]

In this study, an inverse synthetic aperture radar (ISAR) image resolution enhancement algorithm based on joint dictionary learning is proposed, by which two special sets of sparse signals called dictionaries are solved by exploiting numerous high-resolution (HR) and low-resolution (LR) ISAR images. Herein a new coupled dictionary learning algorithm based on restricted Boltzmann machine (RBM) is designed to learn a LR and a HR dictionary using LR and HR image patches. Since the echoes are equivalent to similar scattering-centre models when an object is illuminated by radar signals with same centre frequency and different bandwidth, respectively, it is reasonable to assume the object's LR ISAR image shares the same sparse representation coefficients with its HR ISAR image. When a LR ISAR image is represented sparsely with a LR dictionary, a HR ISAR images can be reconstructed based on a HR dictionary owing to the similar sparse representation coefficients. Experiment results with simulation data demonstrate the superior performance of the proposed method over other classical dictionary training algorithms. [ABSTRACT FROM AUTHOR]

Restricted Boltzmann Machines (RBMs) have been used in a number of applications, but only a few works have addressed them in the context of information security. However, such models have their performance severely affected by some hyperparameters that are usually hand‐tuned. In this work, the authors consider learning features in an unsupervised fashion by means of RBMs fine‐tuned by hypercomplex‐based metaheuristic techniques in the context of malicious content detection. Experiments are conducted over three public datasets and six metaheuristic techniques, which are used to fine‐tune RBM hyperparameters such that RBM extracts features that best represent malicious content present in spam e‐mail messages, and generates a dataset to be used as input to classification through the Optimum Path Forest supervised algorithm. Experimental results demonstrate that a small number of features generated through RBM can achieve a competitive accuracy in relation to the original dataset, however, with lower computational cost. Furthermore, this study presents the power of quaternions for RBMs parameter optimisation, comparing it against the well‐known Harmonic Search, as well as its variants Improved Harmonic Search and Parameter Setting‐Free Harmonic Search. It was concluded that RBM‐based learning techniques are suitable for features extraction in the context of this work. [ABSTRACT FROM AUTHOR]

Boltzmann machines have useful roles in deep learning applications, such as generative data modeling, initializing weights for other types of networks, or extracting efficient representations from high-dimensional data. Most Boltzmann machines use restricted topologies that exclude looping connectivity, as such connectivity creates complex distributions that are difficult to sample. We have used an open-system quantum annealer to sample from complex distributions and implement Boltzmann machines with looping connectivity. Further, we have created policies mapping Boltzmann machine variables to the quantum bits of an annealer. These policies, based on correlation and entropy metrics, dynamically reconfigure the topology of Boltzmann machines during training and improve performance.

The authors present a model where object segmentation and recognition are connected with a bottom‐up inference and a top‐down generation pathway so that the two models can communicate and cooperate with each other. They are integrated into an energy function and optimised at the same time. To achieve the cooperation, objects are modelled in two aspects: shape and appearance, so that the recognition result could feedback to the segmentation process. Restricted Boltzmann machine is employed to learn the shapes of the objects with corresponding labels and perform object recognition based on the object shapes. Another pathway involved with the appearance knowledge of the objects is also established so that both the shape and appearance information will guide and constrain the evolution of the segmentation developing towards the region of interest, which will further facilitate the performances of both tasks. Experiments demonstrate the effectiveness of the proposed model. [ABSTRACT FROM AUTHOR]

Abstract Scene modeling is very crucial for robots that need to perceive, reason about and manipulate the objects in their environments. In this paper, we adapt and extend Boltzmann Machines (BMs) for contextualized scene modeling. Although there are many models on the subject, ours is the first to bring together objects, relations, and affordances in a highly-capable generative model. For this end, we introduce a hybrid version of BMs where relations and affordances are incorporated with shared, tri-way connections into the model. Moreover, we introduce a dataset for relation estimation and modeling studies. We evaluate our method in comparison with several baselines on object estimation, out-of-context object detection, relation estimation, and affordance estimation tasks. Moreover, to illustrate the generative capability of the model, we show several example scenes that the model is able to generate, and demonstrate the benefits of the model on a humanoid robot. The code and the dataset are publicly made available at: https://github.com/bozcani/COSMO. Highlights • Boltzmann Machines (BMs) are adapted and used for contextualized scene modeling. • Being generative, the adapted BMs can handle and predict any missing information in the scene. • The paper demonstrated those capabilities on object, spatial relations and affordances. • The extensive experiments as well as experiments on a humanoid robot illustrate practical uses of the models. [ABSTRACT FROM AUTHOR]

It was recently proposed to leverage the representational power of artificial neural networks, in particular Restricted Boltzmann Machines, in order to model complex quantum states of many-body systems [G. Carleo and M. Troyer, Science355(6325) (2017) 602.]. States represented in this way, called Neural Network States (NNSs), were shown to display interesting properties like the ability to efficiently capture long-range quantum correlations. However, identifying an optimal neural network representation of a given state might be challenging, and so far this problem has been addressed with stöchastic optimization techniques. In this work, we explore a different direction. We study how the action of elementary quantum operations modifies NNSs. We parametrize a family of many body quantum operations that can be directly applied to states represented by Unrestricted Boltzmann Machines, by just adding hidden nodes and updating the network parameters. We show that this parametrization contains a set of universal quantum gates, from which it follows that the state prepared by any quantum circuit can be expressed as a Neural Network State with a number of hidden nodes that grows linearly with the number of elementary operations in the circuit. This is a powerful representation theorem (which was recently obtained with different methods) but that is not directly useful, since there is no general and efficient way to extract information from this unrestricted description of quantum states. To circumvent this problem, we propose a step-wise procedure based on the projection of Unrestricted quantum states to Restricted quantum states. In turn, two approximate methods to perform this projection are discussed. In this way, we show that it is in principle possible to approximately optimize or evolve Neural Network States without relying on stochastic methods such as Variational Monte Carlo, which are computationally expensive. [ABSTRACT FROM AUTHOR]

The background subtraction is an important technique in computer vision which segments moving objects into video sequences by comparing each new frame with a learned background model. In this work, the authors propose a novel background subtraction method based on Gaussian–Bernoulli restricted Boltzmann machines (GRBMs). The GRBM is different from the ordinary restricted Boltzmann machine (RBM) by using real numbers as inputs, resulting in a constrained mixture of Gaussians, which is one of the most widely used techniques to solve the background subtraction problem. The GRBM makes it easy to learn the variance of pixel values and takes the advantage of the generative model paradigm of the RBM. They present a simple technique to reconstruct the learned background model from a given input frame and to extract the foreground from the background using the variance learned for each pixel. Furthermore, they demonstrate the effectiveness of the proposed technique with extensive experimentation and quantitative evaluation on several commonly used public data sets for background subtraction. [ABSTRACT FROM AUTHOR]

In this work we introduce a multi-species generalization of the Hopfield model for associative memory, where neurons are divided into groups and both inter-groups and intra-groups pair-wise interactions are considered, with different intensities. Thus, this system contains two of the main ingredients of modern deep neural-network architectures: Hebbian interactions to store patterns of information and multiple layers coding different levels of correlations. The model is completely solvable in the low-load regime with a suitable generalization of the Hamilton-Jacobi technique, despite the Hamiltonian can be a non-definite quadratic form of the Mattis magnetizations. The family of multi-species Hopfield model includes, as special cases, the 3-layers Restricted Boltzmann Machine with Gaussian hidden layer and the Bidirectional Associative Memory model. [ABSTRACT FROM AUTHOR]

Statistical speech reconstruction for larynx-related dysphonia has achieved good performance using Gaussian mixture models and, more recently, restricted Boltzmann machine arrays; however, deep neural network (DNN)-based systems have been hampered by the limited amount of training data available from individual voice-loss patients. The authors propose a novel DNN structure that allows a partially supervised training approach on spectral features from smaller data sets, yielding very good results compared with the current state-of-the-art.

Secure and stable routing is a critical issue in a conventional network environment, as the control and data planes are integrated. An intelligent routing is highly essential to provide an attack free network to its intended users. The programmable software defined network is an emerging network paradigm in which the security policies and the forwarding rules are implemented in the controller via North Bound Interface. This study incorporates a cognitive neural network learning algorithm namely restricted Boltzmann machine (RBM), to detect the routing‐based distributed denial of service attacks in dynamic source routing (DSR) protocol. RBM is a stochastic and unsupervised learning algorithm that self‐learns the network conditions by using its reasoning capability and segregates malicious routes in the route cache using context‐aware trust metrics such as reputation and energy consumption implemented in North Bound Application Programming Interface. The results show that a cognitive DSR protocol provides secure routing by increased packet delivery ratio, decreased end‐to‐end delay, reduced energy consumption and accurate detection of malicious routes compared with the conventional DSR protocol. [ABSTRACT FROM AUTHOR]

Analyzing quantum many-body problems and elucidating the entangled structure of quantum states is a significant challenge common to a wide range of fields. Recently, a novel approach using machine learning was introduced to address this challenge. The idea is to 'embed' nontrivial quantum correlations (quantum entanglement) into artificial neural networks. Through intensive developments, artificial neural network methods are becoming new powerful tools for analyzing quantum many-body problems. Among various artificial neural networks, this topical review focuses on Boltzmann machines and provides an overview of recent developments and applications., (© 2023 IOP Publishing Ltd.)