Data representations in low dimensions such as results from unsupervised dimensionality reduction methods are often visually interpreted to find clusters of observations. To identify clusters the result must be appreciably clustered. This property of a result may be called "clusteredness". When judged visually, the appreciation of clusteredness is highly subjective. In this paper we suggest an objective way to assess clusteredness in data representations. We provide a definition of clusteredness that captures important aspects of a clustered appearance. We characterize these aspects and define the extremes rigorously. For this characterization of clusteredness we suggest an index to assess the degree of clusteredness, coined the OPTICS Cordillera. It makes only weak assumptions and is a property of the result, invariant for different partitionings or cluster assignments. We provide bounds and a normalization for the index, and prove that it represents the aspects of clusteredness. Our index is parsimonious with respect to mandatory parameters but also exible by allowing optional parameters to be tuned. The index can be used as a descriptive goodness-of-clusteredness statistic or to compare different results. For illustration we use a data set of handwritten digits which are very differently represented in two dimensions by various popular dimensionality reduction results. Empirically, observers had a hard time to visually judge the clusteredness in these representations but our index provides a clear and easy characterisation of the clusteredness of each result. (authors' abstract) / Series: Discussion Paper Series / Center for Empirical Research Methods

Tuberculosis (TB) is an infectious disease which is a global public health problem with over 9 million new cases annually. Tuberculosis treatment, with patient supervision and support is an element of the global plan to stop TB designed by the World Health Organization in 2006. The plan requires prediction of patient treatment course destination. The prediction outcome can be used to determine how intensive the level of supplying services and supports in frame of DOTS therapy should be. No predictive model for the outcome has been developed yet and only limited reports of influential factors for considered outcome are available. To fill this gap, this thesis develops a machine learning approach to predict the outcome of tuberculosis treatment course, which includes, firstly, data of 6,450 Iranian TB patients under DOTS (directly observed treatment, short course ) therapy were analysed to initially diagnose the significant predictors by correlation analysis; secondly, these significant features were applied to find the best classification approach from six examined algorithms including decision tree, Bayesian network, logistic regression, multilayer perceptron, radial basis function, and support vector machine; thirdly, the prediction accuracy of these existing techniques was improved by proposing and developing a new integrated method of k-mean clustering and classification algorithms. Finally, a cluster-based simplified decision tree (CSDT) was developed through an innovative hierarchical clustering and classification algorithm. CSDT was built by k-mean partitioning and the decision tree learning. This innovative method not only improves the prediction accuracy significantly but also leads to a much simpler and interpretative decision tree. The main results of this study included, firstly, finding seventeen significantly correlated features which were: age, sex, weight, nationality, area of residency, current stay in prison, low body weight, TB type, treatment category, length of disease, TB case type, recent TB infection, diabetic or HIV positive, and social risk factors like history of imprisonment, IV drug usage, and unprotected sex ; secondly, the results by applying and comparing six applied supervised machine learning tools on the testing set revealed that decision trees gave the best prediction accuracy (74.21%) compared with other methods; thirdly, by using testing set, the new integrated approach to combine the clustering and classification approach leads to the prediction accuracy improvement for all applied classifiers; the most and least improvement for prediction accuracy were shown by logistic regression (10%) and support vector machine (4%) respectively. Finally, by applying the proposed and developed CSDT, cluster-based simplified decision trees were optioned, which reduced the size of the resulting decision tree and further improved the prediction accuracy. Data type and having normal distribution have created an opportunity for the decision tree to outperform other algorithms. Pre-learning by k-mean clustering to relocate the objects and put similar cases in the same group can improve the classification accuracy. The compatible feature of k-mean partitioning and decision tree to generate pure local regions can simplify the decision trees and make them more precise through creating smaller sub-trees with fewer misclassified cases. The extracted rules from these trees can play the role of a knowledge base for a decision support system in further studies.

Tornado outbreaks (TO) occur every year across the continental United States and are a result of various synoptic scale, mesoscale, and climatological patterns. This study looks to find what patterns exist among the various scales and how that relates to the climatology of the TOs. In order to find these patterns, principal component analysis (PCA) and a cluster analysis were conducted to differentiate the patterns of data. Four distinct clusters of TOs were found with varying synoptic and mesoscale patterns as well as distinct climatological patterns. An interesting result from this study includes the shifting of TO characteristics over time to a more synoptically forced pattern that has becoming stronger and shifted eastward from the Great Plains.

We present a novel approach for vehicle detection and tracking in highway surveillance videos. This method incorporates well-studied computer vision and machine learning techniques to form an unsupervised system, where vehicles are automatically "learned" from video sequences. First an enhanced adaptive background mixture model is used to identify positive and negative examples. Then a video-specific classifier is trained with these examples. Both the background model and the trained classifier are used in conjunction to detect vehicles in a frame. Tracking is achieved by a simplified multi-hypotheses approach. An over-complete set of tracks is created considering every observation within a time interval. As needed hypothesized detections are generated to force continuous tracks. Finally, a scoring function is used to separate the valid tracks in the over-complete set. The proposed detection and tracking algorithm is tested in a challenging application; vehicle counting. Our method achieved very accurate results in three traffic surveillance videos that are significantly different in terms of view-point, quality and clutter. / text

The aim of this work is to introduce modular processing mechanisms for cortical functions implemented in networks of spiking neurons. Neural maps are a feature of cortical processing found to be generic throughout sensory cortical areas, and self-organisation to the fundamental properties of input spike trains has been shown to be an important property of cortical organisation. Additionally, oscillatory behaviour, temporal coding of information, and learning through spike timing dependent plasticity are all frequently observed in the cortex. The traditional self-organising map (SOM) algorithm attempts to capture the computational properties of this cortical self-organisation in a neural network. As such, a cognitive module for a spiking SOM using oscillations, phasic coding and STDP has been implemented. This model is capable of mapping to distributions of input data in a manner consistent with the traditional SOM algorithm, and of categorising generic input data sets. Higher-level cortical processing areas appear to feature a hierarchical category structure that is founded on a feature-based object representation. The spiking SOM model is therefore extended to facilitate input patterns in the form of sets of binary feature-object relations, such as those seen in the field of formal concept analysis. It is demonstrated that this extended model is capable of learning to represent the hierarchical conceptual structure of an input data set using the existing learning scheme. Furthermore, manipulations of network parameters allow the level of hierarchy used for either learning or recall to be adjusted, and the network is capable of learning comparable representations when trained with incomplete input patterns. Together these two modules provide related approaches to the generation of both topographic mapping and hierarchical representation of input spaces that can be potentially combined and used as the basis for advanced spiking neuron models of the learning of complex representations.

Natural images arise from complicated processes involving many factors of variation. They reflect the wealth of shapes and appearances of objects in our three-dimensional world, but they are also affected by factors such as distortions due to perspective, occlusions, and illumination, giving rise to structure with regularities at many different levels. Prior knowledge about these regularities and suitable representations that allow efficient reasoning about the properties of a visual scene are important for many image processing and computer vision tasks. This thesis focuses on models of image structure at intermediate levels of complexity as required, for instance, for image inpainting or segmentation. It aims at developing generative, probabilistic models of this kind of structure, and, in particular, at devising strategies for learning such models in a largely unsupervised manner from data. One hallmark of natural images is that they can often be decomposed into regions with very different visual characteristics. The main approach of this thesis is therefore to represent images in terms of regions that are characterized by their shapes and appearances, and an image is then composed from many such regions. We explore approaches to learn about the appearance of regions, to learn about region shapes, and ways to combine several regions to form a full image. To achieve this goal, we make use of some ideas for unsupervised learning developed in the literature on models of low-level image structure and in the “deep learning” literature. These models are used as building blocks of more structured model formulations that incorporate additional prior knowledge of how images are formed. The thesis makes the following contributions: Firstly, we investigate a popular, MRF based prior of natural image structure, the Field-of Experts, with respect to its ability to model image textures, and propose an extended formulation that is considerably more successful at this task. This formulation gives rise to a fully parametric, translation-invariant probabilistic generative model of image textures. We illustrate how this model can be used as a component of a more comprehensive model of images comprising multiple textured regions. Secondly, we develop a model of region shape. This work is an extension of the “Masked Restricted Boltzmann Machine” proposed by Le Roux et al. (2011) and it allows explicit reasoning about the independent shapes and relative depths of occluding objects. We develop an inference and unsupervised learning scheme and demonstrate how this shape model, in combination with the masked RBM gives rise to a good model of natural image patches. Finally, we demonstrate how this model of region shape can be extended to model shapes in large images. The result is a generative model of large images which are formed by composition from many small, partially overlapping and occluding objects.

"This thesis focuses on two key problems in reinforcement learning: How to design reward functions to obtain intended behaviors in autonomous systems using the learning-based control? Given complex mission specification, how to shape the reward function to achieve fast convergence and reduce sample complexity while learning the optimal policy? To answer these questions, the first part of this thesis investigates inverse reinforcement learning (IRL) method with a purpose of learning a reward function from expert demonstrations. However, existing algorithms often assume that the expert demonstrations are generated by the same reward function. Such an assumption may be invalid as one may need to aggregate data from multiple experts to obtain a sufficient set of demonstrations. In the first and the major part of the thesis, we develop a novel method, called Non-parametric Behavior Clustering IRL. This algorithm allows one to simultaneously cluster behaviors while learning their reward functions from demonstrations that are generated from more than one expert/behavior. Our approach is built upon the expectation-maximization formulation and non-parametric clustering in the IRL setting. We apply the algorithm to learn, from driving demonstrations, multiple driver behaviors (e.g., aggressive vs. evasive driving behaviors). In the second task, we study whether reinforcement learning can be used to generate complex behaviors specified in formal logic — Linear Temporal Logic (LTL). Such LTL tasks may specify temporally extended goals, safety, surveillance, and reactive behaviors in a dynamic environment. We introduce reward shaping under LTL constraints to improve the rate of convergence in learning the optimal and probably correct policies. Our approach exploits the relation between reward shaping and actor-critic methods for speeding up the convergence and, as a consequence, reducing training samples. We integrate compositional reasoning in formal methods with actor-critic reinforcement learning algorithms to initialize a heuristic value function for reward shaping. This initialization can direct the agent towards efficient planning subject to more complex behavior specifications in LTL. The investigation takes the initial step to integrate machine learning with formal methods and contributes to building highly autonomous and self-adaptive robots under complex missions."

Zero-resource speech processing is a growing research area which aims to develop methods that can discover linguistic structure and representations directly from unlabelled speech audio. Such unsupervised methods would allow speech technology to be developed in settings where transcriptions, pronunciation dictionaries, and text for language modelling are not available. Similar methods are required for cognitive models of language acquisition in human infants, and for developing robotic applications that are able to automatically learn language in a novel linguistic environment. There are two central problems in zero-resource speech processing: (i) finding frame-level feature representations which make it easier to discriminate between linguistic units (phones or words), and (ii) segmenting and clustering unlabelled speech into meaningful units. The claim of this thesis is that both top-down modelling (using knowledge of higher-level units to to learn, discover and gain insight into their lower-level constituents) as well as bottom-up modelling (piecing together lower-level features to give rise to more complex higher-level structures) are advantageous in tackling these two problems. The thesis is divided into three parts. The first part introduces a new autoencoder-like deep neural network for unsupervised frame-level representation learning. This correspondence autoencoder (cAE) uses weak top-down supervision from an unsupervised term discovery system that identifies noisy word-like terms in unlabelled speech data. In an intrinsic evaluation of frame-level representations, the cAE outperforms several state-of-the-art bottom-up and top-down approaches, achieving a relative improvement of more than 60% over the previous best system. This shows that the cAE is particularly effective in using top-down knowledge of longer-spanning patterns in the data; at the same time, we find that the cAE is only able to learn useful representations when it is initialized using bottom-up pretraining on a large set of unlabelled speech. The second part of the thesis presents a novel unsupervised segmental Bayesian model that segments unlabelled speech data and clusters the segments into hypothesized word groupings. The result is a complete unsupervised tokenization of the input speech in terms of discovered word types|the system essentially performs unsupervised speech recognition. In this approach, a potential word segment (of arbitrary length) is embedded in a fixed-dimensional vector space. The model, implemented as a Gibbs sampler, then builds a whole-word acoustic model in this embedding space while jointly performing segmentation. We first evaluate the approach in a small-vocabulary multi-speaker connected digit recognition task, where we report unsupervised word error rates (WER) by mapping the unsupervised decoded output to ground truth transcriptions. The model achieves around 20% WER, outperforming a previous HMM-based system by about 10% absolute. To achieve this performance, the acoustic word embedding function (which maps variable-duration segments to single vectors) is refined in a top-down manner by using terms discovered by the model in an outer loop of segmentation. The third and final part of the study extends the small-vocabulary system in order to handle larger vocabularies in conversational speech data. To our knowledge, this is the first full-coverage segmentation and clustering system that is applied to large-vocabulary multi-speaker data. To improve efficiency, the system incorporates a bottom-up syllable boundary detection method to eliminate unlikely word boundaries. We compare the system on English and Xitsonga datasets to several state-of-the-art baselines. We show that by imposing a consistent top-down segmentation while also using bottom-up knowledge from detected syllable boundaries, both single-speaker and multi-speaker versions of our system outperform a purely bottom-up single-speaker syllable-based approach. We also show that the discovered clusters can be made less speaker- and gender-specific by using features from the cAE (which incorporates both top-down and bottom-up learning). The system's discovered clusters are still less pure than those of two multi-speaker unsupervised term discovery systems, but provide far greater coverage. In summary, the different models and systems presented in this thesis show that both top-down and bottom-up modelling can improve representation learning, segmentation and clustering of unlabelled speech data.

This thesis proposes and tests a recursive, adpative, and computationally inexpensive method for partitioning real-number spaces. When tested for proof-of-concept on both one- and two- dimensional classification and control problems, the Fern algorithm was found to work well in one dimension, moderately well for two-dimensional classification, and not at all for two-dimensional control. Testing ferns as pure discretizers - which would involve a secondary discrete learner - has been left to future work. / text

Many high-dimensional time-varying signals can be modeled as a sequence of noisy nonlinear observations of a low-dimensional dynamical process. Given high-dimensional observations and a distribution describing the dynamical process, we present a computationally inexpensive approximate algorithm for estimating the inverse of this mapping. Once this mapping is learned, we can invert it to construct a generative model for the signals. Our algorithm can be thought of as learning a manifold of images by taking into account the dynamics underlying the low-dimensional representation of these images. It also serves as a nonlinear system identification procedure that estimates the inverse of the observation function in nonlinear dynamic system. Our algorithm reduces to a generalized eigenvalue problem, so it does not suffer from the computational or local minimum issues traditionally associated with nonlinear system identification, allowing us to apply it to the problem of learning generative models for video sequences.