Showing papers on "Hidden Markov model published in 2020"

Quantized Nonstationary Filtering of Networked Markov Switching RSNSs: A Multiple Hierarchical Structure Strategy

Abstract: This article addresses the quantized nonstationary filtering problem for networked Markov switching repeated scalar nonlinear systems (MSRSNSs). A more general issue is explored for MSRSNSs, where the measurement outputs are characterized by packet losses, nonstationary quantized output, and randomly occurred sensor nonlinearities (ROSNs) simultaneously. Note that both packet losses and ROSNSs are depicted by Bernoulli distributed sequences. By utilizing a multiple hierarchical structure strategy, the nonstationary filters are designed for MSRSNSs, in which the correlation among modes of systems, quantizer, and controller are presented in terms of nonstationary Markov process. A practical example is provided to verify the proposed theoretical results.

Hidden Markov Model-Based Nonfragile State Estimation of Switched Neural Network With Probabilistic Quantized Outputs

Abstract: This paper focuses on the state estimator design problem for a switched neural network (SNN) with probabilistic quantized outputs, where the switching process is governed by a sojourn probability. It is assumed that both packet dropouts and signal quantization exist in communication channels. Asynchronous estimator and quantification function are described by two different hidden Markov model between the SNNs and its estimator. To deal with the small uncertain of estimators in a random way, a probabilistic nonfragile state estimator is introduced, where uncertain information is described by the interval type of gain variation. A sufficient condition on mean square stable of the estimation error system is obtained and then the desired estimator is designed. Finally, a simulation result is provided to verify the effectiveness of the proposed design method.

A Probabilistic Formulation of Unsupervised Text Style Transfer

Abstract: We present a deep generative model for unsupervised text style transfer that unifies previously proposed non-generative techniques. Our probabilistic approach models non-parallel data from two domains as a partially observed parallel corpus. By hypothesizing a parallel latent sequence that generates each observed sequence, our model learns to transform sequences from one domain to another in a completely unsupervised fashion. In contrast with traditional generative sequence models (e.g. the HMM), our model makes few assumptions about the data it generates: it uses a recurrent language model as a prior and an encoder-decoder as a transduction distribution. While computation of marginal data likelihood is intractable in this model class, we show that amortized variational inference admits a practical surrogate. Further, by drawing connections between our variational objective and other recent unsupervised style transfer and machine translation techniques, we show how our probabilistic view can unify some known non-generative objectives such as backtranslation and adversarial loss. Finally, we demonstrate the effectiveness of our method on a wide range of unsupervised style transfer tasks, including sentiment transfer, formality transfer, word decipherment, author imitation, and related language translation. Across all style transfer tasks, our approach yields substantial gains over state-of-the-art non-generative baselines, including the state-of-the-art unsupervised machine translation techniques that our approach generalizes. Further, we conduct experiments on a standard unsupervised machine translation task and find that our unified approach matches the current state-of-the-art.

Weakly Supervised Sequence Tagging from Noisy Rules

Abstract: We propose a framework for training sequence tagging models with weak supervision consisting of multiple heuristic rules of unknown accuracy. In addition to supporting rules that vote on tags in the output sequence, we introduce a new type of weak supervision, called linking rules, that vote on how sequence elements should be grouped into spans with the same tag. These rules are an alternative to candidate span generators that require significantly more human effort. To estimate the accuracies of the rules and combine their conflicting outputs into training data, we introduce a new type of generative model, linked hidden Markov models (linked HMMs), and prove they are generically identifiable (up to a tag permutation) without any observed training labels. We find that linked HMMs provide an average 7 F1 point boost on benchmark named entity recognition tasks versus generative models that assume the tags are i.i.d. Further, neural sequence taggers trained with these structure-aware generative models outperform comparable state-of-the-art approaches to weak supervision by an average of 2.6 F1 points.

Real-Time Multistep Attack Prediction Based on Hidden Markov Models

Abstract: A novel method based on the Hidden Markov Model is proposed to predict multistep attacks using IDS alerts. We consider the hidden states as similar phases of a particular type of attack. As a result, it can be easily adapted to multistep attacks and foresee the next steps of an attacker. To achieve this goal, a preliminary off-line training phase based on observations will be required. These observations are obtained by matching the IDS alert information with a database previously built for this purpose using a clusterization method from the CVE global database to avoid overfitting. The training model is performed using both unsupervised and supervised algorithms. Once the training is completed and probability matrices are computed, the prediction module compute the best state sequence based on the state probability for each step of the multistep attack in progress using the Viterbi and forward-backward algorithms. The training model includes the mean number of alerts and the number of alerts in progress to assist in obtaining the final attack probability. The model is analyzed for DDoS phases because it is a great problem in all Internet services. The proposed method is validated into a virtual DDoS scenario using current vulnerabilities. The results proving the system's ability to perform real-time prediction.

Dissipativity-Based Asynchronous Fuzzy Sliding Mode Control for T–S Fuzzy Hidden Markov Jump Systems

Abstract: This paper investigates the problem of dissipativity-based asynchronous fuzzy integral sliding mode control (AFISMC) for nonlinear Markov jump systems represented by Takagi–Sugeno (T–S) models, which are subject to external noise and matched uncertainties. Since modes of original systems cannot be directly obtained, the hidden Markov model is employed to detect mode information. With the detected mode and the parallel distributed compensation approach, a suitable fuzzy integral sliding surface is devised. Then using Lyapunov function, a sufficient condition for the existence of sliding mode controller gains is developed, which can also ensure the stochastic stability of the sliding mode dynamics with a satisfactory dissipative performance. An AFISMC law is proposed to drive system trajectories into the predetermined sliding mode boundary layer in finite time. For the case with unknown bound of uncertainties, an adaptive AFISMC law is developed as well. The studied T–S fuzzy Markov jump systems involve both continuous-time and discrete-time domains. Finally, some simulation results are presented to demonstrate the applicability and effectiveness of the proposed approaches.

A Double-Step Unscented Kalman Filter and HMM-Based Zero-Velocity Update for Pedestrian Dead Reckoning Using MEMS Sensors

Abstract: In this paper, we propose a novel method for pedestrian dead reckoning (PDR) using microelectromechanical system magnetic, angular rate, and gravity sensors, which includes a double-step unscented Kalman filter (DUKF) and hidden Markov model (HMM)-based zero-velocity-update (ZVU) algorithm. The DUKF divides the measurement updates of the gravity vector and the magnetic field vector into two steps in order to avoid the unwanted correction for the Euler angle error. The HMM-based ZVU algorithm is developed to recognize the ZVU efficiently. Thus, the proposed PDR method can reduce the position drift caused by the heading error and fault zero-velocity measurement. Experimental results demonstrate that the proposed method achieves better yaw estimate, as well as zero-velocity measurement, and obtains more accurate dead-reckoning position than other methods in the literature.

A Hybrid RNN-HMM Approach for Weakly Supervised Temporal Action Segmentation

Abstract: Action recognition has become a rapidly developing research field within the last decade. But with the increasing demand for large scale data, the need of hand annotated data for the training becomes more and more impractical. One way to avoid frame-based human annotation is the use of action order information to learn the respective action classes. In this context, we propose a hierarchical approach to address the problem of weakly supervised learning of human actions from ordered action labels by structuring recognition in a coarse-to-fine manner. Given a set of videos and an ordered list of the occurring actions, the task is to infer start and end frames of the related action classes within the video and to train the respective action classifiers without any need for hand labeled frame boundaries. We address this problem by combining a framewise RNN model with a coarse probabilistic inference. This combination allows for the temporal alignment of long sequences and thus, for an iterative training of both elements. While this system alone already generates good results, we show that the performance can be further improved by approximating the number of subactions to the characteristics of the different action classes as well as by the introduction of a regularizing length prior. The proposed system is evaluated on two benchmark datasets, the Breakfast and the Hollywood extended dataset, showing a competitive performance on various weak learning tasks such as temporal action segmentation and action alignment.

Passivity-Based Control for Hidden Markov Jump Systems With Singular Perturbations and Partially Unknown Probabilities

Abstract: In this article, the passivity-based control problem is addressed for hidden Markov jump systems with singular perturbations and partially unknown probabilities. The hidden Markov model (HMM) with partially unknown probabilities is introduced, where the partially unknown probabilities may exist in either the transition probability matrix of Markov chain, the observation probability matrix of observed signal, or in both of them. Thus, the underlying HMM is more general than the one in some existing works. By using this HMM, some passivity analysis criteria are established for Markov jump singularly perturbed systems with partial unknown probabilities. Based on these criteria, a unified method is presented for the design of controllers to ensure the passivity of the system. A numerical example and an armature controlled dc motor model are given to illustrate the efficiency of the proposed method.

Dynamic event-based reliable dissipative asynchronous control for stochastic Markov jump systems with general conditional probabilities

Abstract: The paper concerns the problem of reliable dissipative asynchronous controller design for a type of stochastic Markov jump systems (MJSs) with general conditional probabilities in the presence of the dynamic event-triggered rule. Since the mode information of the stochastic MJS is impossible to obtain, a reliable asynchronous controller is established such that the phenomena of nonsynchronous modes between the original stochastic MJS and the developed controller is formulated as a hidden Markov model. A novel dynamic event-triggered strategy is constructed to further decrease the data transmission frequency over the communication network. By taking the internal dynamic variable into account, the criteria of stochastically stable with the reliable dissipativity performance for the resulting closed-loop stochastic MJS are provided based on a collection of linear matrix inequalities. Further, the designs of reliable dissipative asynchronous controller and the dynamic event-triggered strategy are developed simultaneously. Lastly, a numerical instance is supplied to elucidate the superiority of proposed method.

Fuzzy-Model-Based Control for Singularly Perturbed Systems with Nonhomogeneous Markov Switching: A Dropout Compensation Strategy

Abstract: This work addresses the asynchronous control for singularly perturbed systems with nonhomogeneous Markov switching approximated by T-S fuzzy models. The transition probabilities of the nonhomogeneous Markov process are supposed to be time-varying and distinguished by of a polytopic set. As distinct from some reported works, to abate the effect of missing packets in unreliable communication network, a novel dropout compensation strategy is constructed, where packet arriving rate is assumed to be uncertain. Meanwhile, to describe the asynchronizations of the dropout compensation strategy and the controller, the hidden Markov models are absorbed. By resorting to the fuzzy-rule-dependent and parameter-dependent Lyapunov-Krasovskii functional, novel sufficient conditions of ${\mathcal{H}_{\infty}}$ performance are formulated and the fuzzy-based asynchronous controller gains are realized. Finally, to testify the efficiency and applicability of the proposed results, a numerical example and a practical tunnel diode circuit system are provided.

Finite-Time ${\mathscr{H}_{\infty}}$ Asynchronous Control for Nonlinear Markov Jump Distributed Parameter Systems via Quantized Fuzzy Output-Feedback Approach

Abstract: This article focuses on the asynchronous output-feedback control design for a class of nonlinear Markov jump distributed parameter systems based on a hidden Markov model. Initially, the considered systems are represented by a Takagi–Sugeno fuzzy model via a sector nonlinearity approach. Furthermore, asynchronous quantizers are introduced to save the limited communication resource in engineering applications. Then, based on the Lyapunov direct method and some inequality techniques, a series of novel stability criteria, which guarantee the finite-time boundedness and ${\mathscr {H}_{\infty } }$ disturbance attenuation performance of the target plants, is established in the form of spatial differential linear matrix inequalities. Finally, a simulation study is provided to verify the viability of the developed approach.

Resilient Asynchronous $H_{\infty}$ Control for Discrete-Time Markov Jump Singularly Perturbed Systems Based on Hidden Markov Model

Abstract: This paper studies the resilient asynchronous ${H} _{\infty }$ control problem for slow sampling discrete-time uncertain Markov jump singularly perturbed systems (SPSs). Compared with slow state variables feedback controller, a new controller is proposed for slow sampling discrete-time SPSs, which has less conservatism. A more realistic situation, i.e., the system modes cannot be directly acquired for controller design, is considered with the help of hidden Markov model (HMM). The goal is to design a resilient asynchronous controller based on HMM such that the closed-loop system is stochastically stable while meeting an expected ${H}_{\infty }$ performance in the presence of random controller gain fluctuation and the system modes hiding for controller. By utilizing matrix inequality techniques and Lyapunov function method, some criteria are established for the existence of the resilient asynchronous controller. The superiority and practicability of the obtained theoretical results are demonstrated by a numerical example and an inverted pendulum system.

Named Entity Recognition without Labelled Data: A Weak Supervision Approach

Abstract: Named Entity Recognition (NER) performance often degrades rapidly when applied to target domains that differ from the texts observed during training. When in-domain labelled data is available, transfer learning techniques can be used to adapt existing NER models to the target domain. But what should one do when there is no hand-labelled data for the target domain? This paper presents a simple but powerful approach to learn NER models in the absence of labelled data through weak supervision. The approach relies on a broad spectrum of labelling functions to automatically annotate texts from the target domain. These annotations are then merged together using a hidden Markov model which captures the varying accuracies and confusions of the labelling functions. A sequence labelling model can finally be trained on the basis of this unified annotation. We evaluate the approach on two English datasets (CoNLL 2003 and news articles from Reuters and Bloomberg) and demonstrate an improvement of about 7 percentage points in entity-level $F_1$ scores compared to an out-of-domain neural NER model.

${H}_{\infty }$ Filtering for Fuzzy Jumping Genetic Regulatory Networks With Round-Robin Protocol: A Hidden-Markov-Model-Based Approach

Abstract: In this article, a method to design the filter for fuzzy jumping genetic regulatory networks is explored. The case when the filters cannot directly utilize the mode information of the plant is taken into account. A hidden Markov model is introduced to address such a problem. Furthermore, a mature scheduling method, namely round-robin protocol, is employed to optimize the data transmission in genetic regulatory networks. On the basis of the fuzzy model approach and the stochastic analysis technique, some novel conditions ensuring the $H_{\infty }$ performance and stochastic stability of the error system are established. The parameters of the filter can be presented via addressing the convex optimization problem. The feasibility of results is finally illustrated by considering a repressilator model subject to stochastic jumping parameters.

Prediction of Taxi Destinations Using a Novel Data Embedding Method and Ensemble Learning

Abstract: The accurate and timely destination prediction of taxis is of great importance for location-based service applications. Over the last few decades, the popularization of vehicle navigation systems has brought the era of big data to the taxi industry. Existing destination prediction approaches are mainly based on various Markov chain models or trip matching ideas, which require geographical information and may encounter the problem of data sparsity. Other machine learning prediction models are still unsatisfactory in providing favorable results. In this paper, first, we propose use of a novel and efficient data embedding method for time-related feature pre-processing. The key idea behind this is to embed the data into a two-dimensional space before feature selection. Second, we propose use of a novel data-driven ensemble learning approach for destination prediction. This approach combines the respective superiorities of support vector regression and deep learning at different segments of the whole trajectory. Our experiments are conducted on two real data sets to demonstrate that the proposed ensemble learning model can get superior performance for taxi destination prediction. Comparisons also confirm the effectiveness of the proposed data embedding method in the deep learning model.

Non-Intrusive Load Disaggregation by Convolutional Neural Network and Multilabel Classification

Abstract: Non-intrusive load monitoring (NILM) is the main method used to monitor the energy footprint of a residential building and disaggregate total electrical usage into appliance-related signals. The most common disaggregation algorithms are based on the Hidden Markov Model, while solutions based on deep neural networks have recently caught the attention of researchers. In this work we address the problem through the recognition of the state of activation of the appliances using a fully convolutional deep neural network, borrowing some techniques used in the semantic segmentation of images and multilabel classification. This approach has allowed obtaining high performances not only in the recognition of the activation state of the domestic appliances but also in the estimation of their consumptions, improving the state of the art for a reference dataset.

Efficient Activity Recognition in Smart Homes Using Delayed Fuzzy Temporal Windows on Binary Sensors

Abstract: Human activity recognition has become an active research field over the past few years due to its wide application in various fields such as health-care, smart home monitoring, and surveillance. Existing approaches for activity recognition in smart homes have achieved promising results. Most of these approaches evaluate real-time recognition of activities using only sensor activations that precede the evaluation time (where the decision is made). However, in several critical situations, such as diagnosing people with dementia, “preceding sensor activations” are not always sufficient to accurately recognize the inhabitant's daily activities in each evaluated time. To improve performance, we propose a method that delays the recognition process in order to include some sensor activations that occur after the point in time where the decision needs to be made. For this, the proposed method uses multiple incremental fuzzy temporal windows to extract features from both preceding and some oncoming sensor activations. The proposed method is evaluated with two temporal deep learning models (convolutional neural network and long short-term memory), on a binary sensor dataset of real daily living activities. The experimental evaluation shows that the proposed method achieves significantly better results than the real-time approach, and that the representation with fuzzy temporal windows enhances performance within deep learning models.

Named Entity Recognition without Labelled Data: A Weak Supervision Approach

Abstract: Named Entity Recognition (NER) performance often degrades rapidly when applied to target domains that differ from the texts observed during training. When in-domain labelled data is available, transfer learning techniques can be used to adapt existing NER models to the target domain. But what should one do when there is no hand-labelled data for the target domain? This paper presents a simple but powerful approach to learn NER models in the absence of labelled data through weak supervision. The approach relies on a broad spectrum of labelling functions to automatically annotate texts from the target domain. These annotations are then merged together using a hidden Markov model which captures the varying accuracies and confusions of the labelling functions. A sequence labelling model can finally be trained on the basis of this unified annotation. We evaluate the approach on two English datasets (CoNLL 2003 and news articles from Reuters and Bloomberg) and demonstrate an improvement of about 7 percentage points in entity-level F1 scores compared to an out-of-domain neural NER model.

DeepArSLR: A Novel Signer-Independent Deep Learning Framework for Isolated Arabic Sign Language Gestures Recognition

Abstract: Hand gesture recognition has attracted the attention of many researchers due to its wide applications in robotics, games, virtual reality, sign language and human-computer interaction. Sign language is a structured form of hand gestures and the most effective communication way among hear-impaired people. Developing an efficient sign language recognition system to recognize dynamic isolated gestures encounters three major challenges, namely, hand segmentation, hand shape feature representation and gesture sequence recognition. Traditional sign language recognition methods utilize color-based hand segmentation algorithms to segment hands, hand-crafted feature extraction for hand shape representation and Hidden Markov Model (HMM) for sequence recognition. In this paper, a novel framework is proposed for signer-independent sign language recognition using multiple deep learning architectures comprising hand semantic segmentation, hand shape feature representation and deep recurrent neural network. The recently developed semantic segmentation method called DeepLabv3+ is trained using a set of pixel-labeled hand images to extract hand regions from each frame of the input video. Then, the extracted hand regions are cropped and scaled to a fixed size to alleviate hand scale variations. Extracting hand shape features is achieved using a single layer Convolutional Self-Organizing Map (CSOM) instead of relying on transfer learning of pre-trained deep convolutional neural networks. The sequence of extracted feature vectors are then recognized using deep Bi-directional Long Short-Term Memory (BiLSTM) recurrent neural network. BiLSTM network contains three BiLSTM layers, one fully connected and softmax layers. The performance of the proposed method is evaluated using a challenging Arabic sign language database containing 23 isolated words captured from three different users. Experimental results show that the performance of proposed framework outperforms with large margin the state-of-the-art methods for signer-independent testing strategy.

Data-Driven Mode Identification and Unsupervised Fault Detection for Nonlinear Multimode Processes

Abstract: In modern plants, industrial processes typically operate under different states to meet the different requirements of high-quality products. Many monitoring models for industrial processes were constructed based on the prior knowledge (the mechanism's model or the process data characteristics) to monitor such processes (called multimode processes). However, obtaining this prior knowledge is difficult in practice. Efficiently monitoring nonlinear multimode processes without any prior knowledge is an open problem that demands further exploration. Since data from different modes follow different distributions while data from the same mode are considered to be sampled from the same distribution, the modes of multimode processes can be uncovered based on the characteristics of the process data. This article proposes using a Dirichlet process Gaussian mixed model to classify the modes of multimode processes based on historical data, and then, determine the mode types of the monitored data. A nonlinear monitoring strategy based on the t-distributed stochastic neighbor embedding is then proposed to achieve nonlinear dimensionality reduction and visualize the data. Finally, a monitoring index that is integrated with support vector data description is constructed for comprehensive monitoring. The proposed nonlinear multimode framework completely realizes data-driven mode identification and unsupervised fault detection without knowing any prior knowledge. The effectiveness and feasibility of the proposed model are demonstrated using data from a simulated wastewater treatment plant.