A nested infinite Gaussian mixture model for identifying known and unknown audio events

Air quality has always been a hot issue of concern to the people, the environmental protection department and the government. Among the massive air quality data, abnormal data can interfere with subsequent experiments and analysis. Therefore, it is necessary to detect abnormal data to improve the accuracy of the data. However, traditional air outlier detection methods require at least one year's data to make inferences about air quality. This paper firstly analyzes the characteristics of air quality big data, and then proposes a framework based on Bayesian non-parametric clustering, namely Dirichlet Process (DP) clustering framework, to realize the outlier detection of air quality. The framework optimizes Gaussian mixture model into infinite Gaussian mixture model according to the results of data analysis, and uses neural network to cluster the data processed by infinite Gaussian mixture model, which effectively improves the clustering accuracy and avoids the need of collecting a large number of training data.

In this work we present and apply a new Hyperspectral image classification method based on Dirichlet Process Gaussian mixture models (also known as infinite Gaussian mixture models - IGMMs). Since this approach is a non-parametric Bayesian method, we circumvent the problem of model selection which is unavoidable and often difficult when employing traditional parametric Gaussian mixture models (GMM). In order to infer model parameters from observations, we adopt a Gibbs Sampling method to sample the posterior distributions of these parameters. In the preprocessing step, we use Local Fisher's Discriminant Analysis (LFDA) for dimension reduction since we expect it to preserve the multi-modal, non-Gaussian structure of the hyperspectral data. We compared our proposed IGMM based classification method to the existing state-of-the-art classification methods using popular hyperspectral imagery datasets. To simulate challenging real-world sensing environment, we mix every pixel with background and study the performance of these classification approaches. Our experiments show that the proposed LFDA-IGMM method and GMM method have almost the same performance for most cases, given “pure” data - with mixed pixels, LFDA-IGMM outperformed GMM for sub-pixel classification. Both methods outperformed the other commonly used classification approaches when the training sample size is relatively large. Also, the IGMM based method is less sensitive to pixel mixing than other methods.

We present a novel generative model for audio event transcription that recognizes “events” on audio signals including multiple kinds of overlapping sounds. In the proposed model, firstly, the overlapping audio events are modeled based on nonnegative matrix factorization into which Bayesian nonparametric approaches: the Markov Indian buffet process and the Chinese restaurant process, are incorporated. This approach allows us to automatically transcribe the events while avoiding the model selection problem by assuming a countably infinite number of possible audio events in the input signal. Then, Bayesian logistic regression annotates the audio frames with the multiple event labels in a semi-supervised learning setup. Experimental results show that our model can better annotate an audio signal in comparison with a baseline method. Additionally, we verify that our infinite generative model is also able to detect unknown audio events that are not included in the training data.

Unsupervised speech processing methods are essential for applications ranging from zero-resource speech technology to modelling child language acquisition. One challenging problem is discovering the word inventory of the language: the lexicon. Lexical clustering is the task of grouping unlabelled acoustic word tokens according to type. We propose a novel lexical clustering model: variable-length word segments are embedded in a fixed-dimensional acoustic space in which clustering is then performed. We evaluate several clustering algorithms and find that the best methods produce clusters with wide variation in sizes, as observed in natural language. The best probabilistic approach is an infinite Gaussian mixture model (IGMM), which automatically chooses the number of clusters. Performance is comparable to that of non-probabilistic Chinese Whispers and average-linkage hierarchical clustering. We conclude that IGMM clustering of fixed-dimensional embeddings holds promise as the lexical clustering component in unsupervised speech processing systems.

For reducing the channel switching frequency due to uncertain and frequent primary users' interruptions, we design a nonparametric channel allocation algorithm which can effectively manage the spectrum resources in distributed cognitive radio networks and serve for delay-sensitive data traffics. In the proposed algorithm, an infinite Gaussian mixture model is used to get the posterior decision from the feature spaces of the channel idle duration and the primary user's frequency channel, while the collapsed Gibbs sampling is applied for fast convergence. Experimental and simulation results demonstrate that the proposed channel allocation algorithm outperforms the random channel allocation method and the idle probability based channel allocation algorithm in terms of throughput and packet drops.

We present a new framework to detect various types of variable objects within massive astronomical time-series data. Assuming that the dominant population of objects is non-variable, we find outliers from this population by using a non-parametric Bayesian clustering algorithm based on an infinite GaussianMixtureModel (GMM) and the Dirichlet Process. The algorithm extracts information from a given dataset, which is described by six variability indices. The GMM uses those variability indices to recover clusters that are described by six-dimensional multivariate Gaussian distributions, allowing our approach to consider the sampling pattern of time-series data, systematic biases, the number of data points for each light curve, and photometric quality. Using the Northern Sky Variability Survey data, we test our approach and prove that the infinite GMM is useful at detecting variable objects, while providing statistical inference estimation that suppresses false detection. The proposed approach will be effective in the exploration of future surveys such as GAIA, Pan-Starrs, and LSST, which will produce massive time-series data.

An effective and non-invasive audio monitoring system needs to be capable of simultaneous real-time detection of multiple audio events in many different environments, and locally executable on resource constrained devices, such as, smart microphones. A major challenge in this research domain is having limited available annotated data. This paper presents a novel framework to generate robust detection models of environmental and human audio events with limited available data. The framework presents the generation of a large synthetic dataset using limited data for any audio event, a novel computationally efficient feature modeling technique, named Audio2Vec, that is robust against environmental variations, and identifies and exploits the syntactic relation between audio states represented by the features and the targeted audio events. The presented framework achieves 10.3% higher F-1 scores compared to the best baseline approaches. To demonstrate the effectiveness of the framework we implemented a real-time audio monitoring system that simultaneously detects 10 audio events on a Raspberry Pi 3B and evaluate it in real home and in-car settings, that achieve F-1 scores of 0.96 and 0.956, respectively.

This paper studies convergence behavior of latent mixing measures that arise in finite and infinite mixture models, using transportation distances (i.e., Wasserstein metrics). The relationship between Wasserstein distances on the space of mixing measures and f-divergence functionals such as Hellinger and Kullback-Leibler distances on the space of mixture distributions is investigated in detail using various identifiability conditions. Convergence in Wasserstein metrics for discrete measures implies convergence of individual atoms that provide support for the measures, thereby providing a natural interpretation of convergence of clusters in clustering applications where mixture models are typically employed. Convergence rates of posterior distributions for latent mixing measures are established, for both finite mixtures of multivariate distributions and infinite mixtures based on the Dirichlet process.

Robotic agents that can explore and sample in a completely unsupervised fashion could greatly increase the amount of scientific data gathered in dangerous and inaccessible environments. Our application is imaging the benthos using an autonomous underwater vehicle with limited communication to surface craft. Robotic exploration of this nature demands in situ data analysis. To this end, this paper presents results of using a Gaussian Mixture Model (GMM), a Hidden Markov Model (HMM) filter, an Infinite Gaussian Mixture Model (IGMM) and a Variation Dirichlet Process model (VDP) for the classification of benthic habitats. All of the models are trained using unsupervised methods. Furthermore, the IGMM and VDP are trained without knowing the the number of classes in the dataset. It was found that the sequential information the HMM filter provides to the classification process adds lag to the habitat boundary estimates, reducing the classification accuracy. The VDP proved to be the most accurate classifier of the four tested, and also one of the fastest to train. We conclude that the VDP is a powerful model for entirely autonomous labelling of benthic datasets.

The diagnosis of telecommunication networks remains a challenging task, mainly due to the large variety and volume of data from which the root causes have to be inferred. Expert systems, supervised machine learning, or Bayesian networks require expensive and time consuming data labeling or processing by experts. In this paper, we propose an Infinite Multivariate Categorical Mixture Model for clustering patterns of faults from data gathered from telecommunication networks. The model is able to automatically identify the number of clusters necessary to explain the data using the Dirichlet process prior. We show how to use Variational Inference to derive an Expectation-Maximization (EM) like algorithm to perform inference on the model. We apply our model on synthetic data generated from an expert Bayesian network of a Fiber- To- The-Home (FTTH) Gigabit capable Passive Optical Network (GPON). We show that the model discovers the patterns linked to the root causes of the faults with up to 96 % accuracy in an unsupervised manner. We also apply our method on real data gathered from the FTTH network and the local area network and demonstrate how the model is able to identify known faults.

Classification is one of the widely used techniques to quantify and mine the rich information present in hyperspectral imagery. However, the realization of trustworthy classification still continues to be a challenging task. This is due to the presence of various uncertainties, such as incomplete a priori knowledge on the actual number of classes and noise. Especially, mismatching between the number of spectral classes and the number of information classes leads to substantial omission or commission errors in the classification. In this letter, we present a new multiclass classification framework, named supervised cascaded classifier system (SC <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> S), that addresses the abovementioned problem by providing reliable results. The SC <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> S method is a two-stage cascaded classification procedure that involves quantifying the uncertainty and then classifying the samples. In the first stage, pixels for which no reliable training samples can be found in the training stage are detected. In the second stage, pixels that are matched with the respective training distribution are classified using a supervised learning algorithm, otherwise labeled as unknown. The proposed SC <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> S framework has been implemented on eight different classification scenarios with a varying number of unknown classes (UCs) using hyperspectral and multispectral imageries. The performance of the proposed SC <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> S method is compared with other widely used classifiers with and without reject-option. The experimental results indicate that our method offers superior classification results even in the case of data sets with a large number of unknown spectral classes.

Lately, the enormous generation of databases in almost every aspect of life has created a great demand for new powerful tools for turning data into useful information. Therefore, researchers were encouraged to explore and develop new machine learning ideas and methods. Mixture models are one of the machine learning techniques receiving considerable attention due to their ability to handle efficiently and effectively multidimensional data. In this paper, we represent a solution for two challenging issues: modeling non-Gaussian data and determining the set of relevant features in the data. The problem of modeling non-Gaussian data largely present in several computer vision, image processing, medical, and Bioinformatics applications is accomplished by the development of a generative infinite Gamma mixture model. The Gamma is chosen for its ability to handle long-tailed distributions, which allows it to have a good approximation to data with outliers. The proposed model, which can be viewed as a Dirichlet process mixture of Gamma distributions, takes into account the feature selection problem by determining a set of relevant features for each data cluster which provides better interpretability and generalization capabilities. We propose then an efficient algorithm to learn this infinite model's parameters by estimating all its posterior quantities of interest using Markov Chain Monte Carlo (MCMC) simulations. Thus, our algorithm is able to perform model selection, parameter learning, and feature selection simultaneously in a single step for the Gamma Mixture model. Furthermore, we show how the model can be used, while comparing it with other popular models in the literature, in two challenging applications namely medical images and gene expressions classification.

For an autonomous mobile robot that works in real-world environment, recognition of its surrounding environment is necessary. This paper presents an object identification system, which can identify known and unknown objects and estimate their locations using 3D point cloud data acquired from a 3D LIDAR sensor mounted on the mobile robot. The proposed system is divided into two main steps; the first step is segmentation based on the 3D point cloud and the second step is identification of the extracted objects. The 3D LIDAR sensor gives sparse 3D shape information accurately, and covers a wide range. One of the main problems of such data is that the object shape information varies according to object's orientation and distance from the sensor. To solve this problem, we use nested infinite Gaussian mixture models for object identification. The experimental results show that the proposed system can extract various types of objects and identify both known and unknown objects.

Co-localization of mobile users combines methods of detecting nearby users and providing them interesting and useful services or information. By exploiting the massive use of smartphones, nearby users can be co-localized using only their captured ambient radio signals. In this paper, we propose a real-time co-localization system, in a centralized manner, that leverages co-located users with high accuracy. We exploit the similarity of radio frequency measurements from users' mobile terminal. We do not require any further information about them. Our co-localization system is based on a nonparametric Bayesian method called infinite Gaussian mixture model that allows the model parameters to change with observed input data. In addition, we propose a modified version of Gibbs sampling technique with an average similarity threshold to better fit user's group. We design our system in a completely centralized manner. Hence, it enables the network to control and manage the formation of the users' groups. We first evaluate the performance of our proposal numerically. Then, we carry out an extensive experiment to demonstrate the feasibility, and the efficiency of our approach with data sets from a real-world setting. Results on experiment favor our algorithm over the state-of-the-art community detection-based clustering method.

The diagnosis of Alzheimer’s disease (AD) faces two important issues. They are how to extract the features of the rhythms of patients with AD and how to label them and reveal the degree of dementia in patients. This study defines 14 instantaneous power indicators of dementia judgment through Hilbert marginal spectrum (HMS) from rhythm waves. A warped infinite Gaussian mixture model (WiGMM) is proposed to learn the latent variables of these indicators to detect the degree of dementia. The experimental results show that HMS-based indicators are able to reflect the cognitive function of AD patients. This proposed method has the ability to detect brain cognitive status through a warped transform and Dirichlet process parameter prior to inference.