A New Approach to Learning in Dynamic Bayesian Networks (DBNs)

Bayesian network models provide an attractive framework for multimodal sensor fusion. They combine an intuitive graphical representation with efficient algorithms for inference and learning. However, the unsupervised nature of standard parameter learning algorithms for Bayesian networks can lead to poor performance in classification tasks. We have developed a supervised learning framework for Bayesian networks, which is based on the Adaboost algorithm of Schapire and Freund. Our framework covers static and dynamic Bayesian networks with both discrete and continuous states. We have tested our framework in the context of a novel multimodal HCI application: a speech-based command and control interface for a Smart Kiosk. We provide experimental evidence for the utility of our boosted learning approach.

Prognostics and health management (PHM) is one of the research hotspots in reliability, where remaining useful life (RUL) estimation is a typical application scenario. In this article, a feature learning method based on dynamic Bayesian networks (DBNs) is proposed to improve the RUL estimation accuracy of the regression models. The best feature set is obtained with the conditional dependencies represented by the DBN structure. A local modeling method is applied here to reduce the computation for high-order DBN construction. The strength of the connections between variables together with a contribution index of variables in the DBN structure are defined to represent the feature importance of the variables. Feature transfer is carried out with feature importance under different operating conditions for a further improvement. Nonlinear regression models such as support vector regression (SVR) and Gaussian mixture regression (GMR) are built based on the learned features and used to estimate the RUL. The turbofan engine dataset C-MAPSS is used to validate the effectiveness of the proposed method. Compared with other recent RUL estimation models, the proposed method has a faster modeling speed and higher prediction accuracy.

Modeling interactions of multiple co-occurring objects in a complex activity is becoming increasingly popular in the video domain. The Dynamic Bayesian Network (DBN) has been applied to this problem in the past due to its natural ability to statistically capture complex temporal dependencies. However, standard DBN structure learning algorithms are generatively learned, require manual structure definitions, and/or are computationally complex or restrictive. We propose a novel structure learning solution that fuses the Granger Causality statistic, a direct measure of temporal dependence, with the Adaboost feature selection algorithm to automatically constrain the temporal links of a DBN in a discriminative manner. This approach enables us to completely define the DBN structure prior to parameter learning, which reduces computational complexity in addition to providing a more descriptive structure. We refer to this modeling approach as the Granger Constraints DBN (GCDBN). Our experiments show how the GCDBN outperforms two of the most relevant state-of-the-art graphical models in complex activity classification on handball video data, surveillance data, and synthetic data.

Fault detection and diagnosis for missing data systems with a three time-slice dynamic Bayesian network approach

Bayesian networks are a concise graphical formalism for describing probabilistic models. We have provided a brief tutorial of methods for learning and inference in dynamic Bayesian networks. In many of the interesting models, beyond the simple linear dynamical system or hidden Markov model, the calculations required for inference are intractable. Two different approaches for handling this intractability are Monte Carlo methods such as Gibbs sampling, and variational methods. An especially promising variational approach is based on exploiting tractable substructures in the Bayesian network.

Using DBN and evidence-based reasoning to develop a risk performance model to interfere ship navigation process safety in Arctic waters

This paper is about the automatic structuring of multiparty meetings using audio information. We have used a corpus of 53 meetings, recorded using a microphone array and lapel microphones for each participant. The task was to segment meetings into a sequence of meeting actions, or phases. We have adopted a statistical approach using dynamic Bayesian networks (DBNs). Two DBN architectures were investigated: a two-level hidden Markov model (HMM) in which the acoustic observations were concatenated; and a multistream DBN in which two separate observation sequences were modelled. Additionally we have also explored the use of counter variables to constrain the number of action transitions. Experimental results indicate that the DBN architectures are an improvement over a simple baseline HMM, with the multistream DBN with counter constraints producing an action error rate of 6%. 1.

An approach to user modelling with discrete stochastic processes is presented, which aims at the dynamic individualization of user interfaces on the syntactic layer. The state of the art of syntactic user modelling is surveyed. The mathematical background of simple Markov Chains and the 'classic' Hidden Markov Model is presented. Furthermore, dynamic Bayesian Networks are introduced, which generalize these Markovian Models. A case study of the simulation experiments uses a multimodal user interface for supervisory control of advanced manufacturing cells. A corresponding simulation model is created and exploited to generate interaction cases, which are the empirical basis for the evaluation. Six topologies of dynamic Bayesian Networks are evaluated for 100 interaction cases and 50 replications each: (1) Markov Chain of order 1, (2) Hidden Markov Model, (3) autoregressive Hidden Markov Model, (4) factorial Hidden Markov Model, (5) simple hierarchical Hidden Markov Model, and (6) tree structured Hidden Markov Model. The dependent variable is the prediction accuracy for a single prediction lead. In a first step, a one-way analysis of variance in conjunction with Tukey's post-hoc test demonstrated a significant superiority of the simple hierarchical Hidden Markov Model. In a second step, an additional two-way analysis of variance also indicated a significantly better prediction accuracy of the simple hierarchical Hidden Markov Model compared to the Hidden Markov Model, but the number of interaction cases also had a significant effect. Hence, the modeller has to take both factors--model topology and number of interaction cases--into account when designing syntactic user models with stochastic processes.

Gait recognition based on improved dynamic Bayesian networks

Buildings are subject to faults in their heating, ventilation and air-conditioning (HVAC) systems that can lead to excessive energy wastage, poor indoor climate, equipment failures and high maintenance costs. Field studies have shown that employing fault detection, diagnosis and prognosis (FDDP) tools followed up with equipment services and corrections can help achieve up to 40% of energy savings within the HVAC system and improve indoor climate, increase equipment lifecycle and reduce maintenance costs. The increasing adoption of building automation systems (BAS), Internet of Things (IoT) and other smart technologies in recent years have allowed large amounts of data to be continuously collected from building systems. This data-rich environment, along with the surge in data analytics and machine learning tools, has made cost-effective data-driven FDDP strategies possible. Compared to purely physics-based methods, data-driven methods require less explicit knowledge of the underlying physical system, thus are often easier to develop, and can learn certain intricate relationships that exist among data. Within the reported data-driven FDDP methods, there exists a few research gaps: 1) data imputation methods that leverage mutual information from correlated measurements to defy poor data quality from BAS have not been utilized efficiently; 2) there lacks a systematic and scalable fault diagnosis framework that incorporates probabilistic temporal relationships to track fault evolution; 3) existing fault diagnosis strategies typically focus on traditional rule-based control strategies and their scalability for advanced control strategies such as Guideline 36 have not been explored yet; 4) active threats information such as cyber-attacks, are typically not incorporated in an FDDP framework; 5) fault prognosis strategies to preemptively identify gradual faults for predictive maintenance have rarely been studied. This research attempts to address the above-mentioned research gaps through the following: Data Imputation: reported data imputation methods that are suitable for handling and repairing multi-source BAS data are evaluated. Data collected from a medium-sized, mixed-use institution building situated in Stockholm, Sweden and a small commercial building simulated in a laboratory setup is used to evaluate five different data imputation methods. Results demonstrate that incorporating time-lagged cross-correlations within the k-Nearest Neighbor (kNN) method helps to significantly improve the imputation accuracy and minimize the impact of repaired data on data-driven algorithms. Dynamic Bayesian Network (DBN)-based Framework for Cyber-Physical Fault Diagnosis: a DBN framework with discretized conditional probabilities parameters to represent the temporal relationships among building measurements is developed. Both domain knowledge and machine learning methods are used to develop the DBN structure and parameter model. The developed framework is evaluated for both traditional rule-based and Guideline 36 controls using datasets from a real building, a laboratory building, and a virtual testbed. Results show that the developed DBN framework is effective in diagnosing and isolating faults in systems even with different control strategies. The framework also successfully distinguishes whether system abnormalities originate from cyber-attacks or naturally occurring physical faults. Potential future direction to improve fault isolation using modified DBN topological structure is also reported in this study. DBN-based Framework for Fault Prognosis: an extension of the DBN framework in conjunction with Robust Multivariate Temporal (RMT) variate selection is proposed for fault prognosis. The RMT variate selection is used to extract localized temporal features from high dimensional datasets to determine the best inputs for training forecasting models. The expected fault-free behavior of multiple target variates, selected using domain knowledge, is forecasted using incoming data. The prediction errors generated from the forecasting phase are used as evidences in the DBN inference to estimate future fault probabilities. Gradual faults simulated in the virtual testbed are used to evaluate the prognosis framework. Results show that the developed framework is effective in prognosing gradual faults by leveraging the trending growth on the prediction errors. The research presented in this thesis contributes to the overall objective of developing a robust and cost-effective DBN-based framework for fault diagnosis and prognosis of building HVAC systems. Potential solutions to other existing challenges of implementing data-driven FDDP strategies, such as obtaining high-quality datasets, handling and repairing missing data, establishing a baseline model for detecting abnormalities despite other disturbances such as weather and internal conditions changes, and extracting temporal features from timeseries data are also examined.

This paper introduces the concept of eigen-dynamics and proposes an eigen dynamics analysis (EDA) method to learn the dynamics of natural hand motion from labelled sets of motion captured with a data glove. The result is parameterized with a high-order stochastic linear dynamic system (LDS) consisting of five lower-order LDS. Each corresponding to one eigen-dynamics. Based on the EDA model, we construct a dynamic Bayesian network (DBN) to analyze the generative process of a image sequence of natural hand motion. Using the DBN, a hand tracking system is implemented. Experiments on both synthesized and real-world data demonstrate the robustness and effectiveness of these techniques.

The following chapter can be understood as one sort of brief introduction to the history and basics of the Hidden Markov Models. Hidden Markov Models (HMMs) are learnable finite stochastic automates. Nowadays, they are considered as a specific form of dynamic Bayesian networks. Dynamic Bayesian networks are based on the theory of Bayes (Bayes & Price, 1763). A Hidden Markov Model consists of two stochastic processes. The first stochastic process is a Markov chain that is characterized by states and transition probabilities. The states of the chain are externally not visible, therefore “hidden”. The second stochastic process produces emissions observable at each moment, depending on a state-dependent probability distribution. It is important to notice that the denomination “hidden” while defining a Hidden Markov Model is referred to the states of the Markov chain, not to the parameters of the model. The history of the HMMs consists of two parts. On the one hand there is the history of Markov process and Markov chains, and on the other hand there is the history of algorithms needed to develop Hidden Markov Models in order to solve problems in the modern applied sciences by using for example a computer or similar electronic devices.

Background: Predicting potential drug resistance mutations are important when evaluating protein-drug interactions of potential new antiviral drugs. Here we used evolutionary data from the Retroviral Aspartyl Protease (RVP) family (PF00077, 54135 sequences) to estimate plausible PI resistant-associated AAS within the H1P. Methods: Using a Hidden Markov Model (HMM) of the RVP family probabilities were extracted for each possible AAS limited to the 38 positions reported in the IAS drug resistance listing for H1P (December 2008 version). The HMM is a dynamic Bayesian network, modeling sequences of amino acids. The HMM is based on curated and representative sequences from the RVP family. Results: Theoretically 760 AAS (20 × 38) are possible for the 38 evaluated positions within the H1P. Of these, the RVP-HMM detected a total of 229 AAS (30.1%) with a probability above 1/20 (0.05). Of the 229 AAS, 51 (70%) were among the 73 AAS included in the IAS listing as PI-resistant mutations, leaving 178 AAS with P > 0.05 as evolutionary plausible. Conclusion: Based on exploration of the RVP family by HMM, 70% of the established PI-resistant associated AAS could be predicted to occur. Additional 178 AAS was identified as evolutionary plausible and potentially could allow for drug-resistance. In conclusion, we provide a probability landscape of plausible/unfavorable AAS based on inherited structure through evolution and genetic distance, which could prove useful for future drug design.

In this work, ail experimental study of the coupled hidden Markov models (CHMMs) is carried out. CHMMs are directed graphical models of stochastic processes and are a special type of Dynamic Bayesian Networks (DBNs). The DBNs generalize the hidden Markov models by representing the hidden states as state variables, and allow the states to have complex interdependencies. This study considers the probabilistic inference problem and the learning problem of these models. A series of experiments were carried out to evaluate the relationship between the learning outcome and various factors in the learning process. In addition, this study looks into the capabilities of the CHMMs in the context of Maximum Likelihood classification of temporal patterns. Empirical results suggest that even with perfect learning, the classification error can be significant in some cases, and it is important to limit the state space of the models when considering the framework in real-world applications.

In this work, we investigate the combination of PGM (Propabilistic Graphical Models) classifiers, either independent or coupled, for the recognition of Arabic handwritten words. The independent classifiers are vertical and horizontal HMMs (Hidden Markov Models) whose observable outputs are features extracted from the image columns and the image rows respectively. The coupled classifiers associate the vertical and horizontal observation streams into a single DBN (Dynamic Bayesian Network). A novel method to extract word baseline and a simple and easily extractable features to construct feature vectors for words in the vocabulary are proposed. Some of these features are statistical, based on pixel distributions and local pixel configurations. Others are structural, based on the presence of ascenders, descenders, loops and diacritic points. Experiments on handwritten Arabic words from IFN/ENIT strongly support the feasibility of the proposed approach. The recognition rates achieve 90.42% with vertical and horizontal HMM, 85.03% and 85.21% with respectively a first and a second DBN which outperform results of some works based on PGMs.