Data visualisation and exploration with prior knowledge

Visualising data for exploratory analysis is a major challenge in many applications. Visualisation allows scientists to gain insight into the structure and distribution of the data, for example finding common patterns and relationships between samples as well as variables. Typically, visualisation methods like principal component analysis and multi-dimensional scaling are employed. These methods are favoured because of their simplicity, but they cannot cope with missing data and it is difficult to incorporate prior knowledge about properties of the variable space into the analysis; this is particularly important in the high-dimensional, sparse datasets typical in geochemistry. In this paper we show how to utilise a block-structured correlation matrix using a modification of a well known non-linear probabilistic visualisation model, the Generative Topographic Mapping (GTM), which can cope with missing data. The block structure supports direct modelling of strongly correlated variables. We show that including prior structural information it is possible to improve both the data visualisation and the model fit. These benefits are demonstrated on artificial data as well as a real geochemical dataset used for oil exploration, where the proposed modifications improved the missing data imputation results by 3 to 13%.

Risk assessment applied to comsumer products with reference to CE marking machines for use at work

New Approach’ Directives now govern the health and safety of most products whether destined for workplace or domestic use. These Directives have been enacted into UK law by various specific legislation principally relating to work equipment, machinery and consumer products. This research investigates whether the risk assessment approach used to ensure the safety of machinery may be applied to consumer products. Crucially, consumer products are subject to the Consumer Protection Act (CPA) 1987, where there is no direct reference to “assessing risk”. This contrasts with the law governing the safety of products used in the workplace, where risk assessment underpins the approach. New Approach Directives are supported by European harmonised standards, and in the case of machinery, further supported by the risk assessment standard, EN 1050. The system regulating consumer product safety is discussed, its key elements identified and a graphical model produced. This model incorporates such matters as conformity assessment, the system of regulation, near miss and accident reporting. A key finding of the research is that New Approach Directives have a common feature of specifying essential performance requirements that provide a hazard prompt-list that can form the basis for a risk assessment (the hazard identification stage). Drawing upon 272 prosecution cases, and with thirty examples examined in detail, this research provides evidence that despite the high degree of regulation, unsafe consumer products still find their way onto the market. The research presents a number of risk assessment tools to help Trading Standards Officers (TSOs) prioritise their work at the initial inspection stage when dealing with subsequent enforcement action.

Finding structures in information networks using the affinity network

This thesis proposes a novel graphical model for inference called the Affinity Network,which displays the closeness between pairs of variables and is an alternative to Bayesian Networks and Dependency Networks. The Affinity Network shares some similarities with Bayesian Networks and Dependency Networks but avoids their heuristic and stochastic graph construction algorithms by using a message passing scheme. A comparison with the above two instances of graphical models is given for sparse discrete and continuous medical data and data taken from the UCI machine learning repository. The experimental study reveals that the Affinity Network graphs tend to be more accurate on the basis of an exhaustive search with the small datasets. Moreover, the graph construction algorithm is faster than the other two methods with huge datasets. The Affinity Network is also applied to data produced by a synchronised system. A detailed analysis and numerical investigation into this dynamical system is provided and it is shown that the Affinity Network can be used to characterise its emergent behaviour even in the presence of noise.

The role of disturbance and resource availability in structuring plant communities

Disturbances alter competitive hierarchies by reducing populations and altering resource regimes. The interaction between disturbance and resource availability may strongly influence the structure of plant communities, as observed in the recolonization of seagrass beds in outer Florida Bay that were denuded by sea-urchin overgrazing. There is no consensus concerning the interaction between disturbance and resource availability on competition intensity (CI). On the other hand, species diversity is dependent on both factors. Peaks in species diversity have been observed to occur when both resource availability and disturbance intensity are high, thus implying that CI is low. Based on this supposition of previous models, I presented the resource-disturbance hypothesis as a graphical model to make predictions of CI as a function of both disturbance intensity and the availability of a limiting resource. The predictions of this model were tested in two experiments within a seagrass community in south Florida, in which transplants of Halodule wrightii were placed into near-monocultures of Syringodium filiforme in a full-factorial array. In the first experiment, two measures of relative CI were calculated based on the changes in the short-shoot number (SS) and of rhizome length (RHL) on the transplants. Both light and disturbance were identified as important factors, though the interaction between light * disturbance was not significant. Relative CISS ranged between 0.2 and 1.0 for the high light and high disturbance treatments and the relative CIRHL < 0 for the same treatments, though results were not significantly different due to high variability and low sample size. These results, including a contour schematic using six data points from the different treatment combinations, preliminarily suggests that the resource-disturbance hypothesis may be used may be used as a next step in developing our understanding of the mechanisms involved in structuring plant communities. Furthermore, the focus of the model is on the outcome of CI, which may be a useful predictor of changes in species diversity. Further study is needed to confirm the results of this study and validate the usefulness of this model in other systems. ^

Modelo para reconstrução 3D de cenas baseado em imagens

3D Reconstruction is the process used to obtain a detailed graphical model in three dimensions that represents some real objectified scene. This process uses sequences of images taken from the scene, so it can automatically extract the information about the depth of feature points. These points are then highlighted using some computational technique on the images that compose the used dataset. Using SURF feature points this work propose a model for obtaining depth information of feature points detected by the system. At the ending, the proposed system extract three important information from the images dataset: the 3D position for feature points; relative rotation and translation matrices between images; the realtion between the baseline for adjacent images and the 3D point accuracy error found.

Karst hazard analysis towards the development of predictive karst models for southern Ontario

An investigation into karst hazard in southern Ontario has been undertaken with the intention of leading to the development of predictive karst models for this region. The reason these are not currently feasible is a lack of sufficient karst data, though this is not entirely due to the lack of karst features. Geophysical data was collected at Lake on the Mountain, Ontario as part of this karst investigation. This data was collected in order to validate the long-standing hypothesis that Lake on the Mountain was formed from a sinkhole collapse. Sub-bottom acoustic profiling data was collected in order to image the lake bottom sediments and bedrock. Vertical bedrock features interpreted as solutionally enlarged fractures were taken as evidence for karst processes on the lake bottom. Additionally, the bedrock topography shows a narrower and more elongated basin than was previously identified, and this also lies parallel to a mapped fault system in the area. This suggests that Lake on the Mountain was formed over a fault zone which also supports the sinkhole hypothesis as it would provide groundwater pathways for karst dissolution to occur. Previous sediment cores suggest that Lake on the Mountain would have formed at some point during the Wisconsinan glaciation with glacial meltwater and glacial loading as potential contributing factors to sinkhole development. A probabilistic karst model for the state of Kentucky, USA, has been generated using the Weights of Evidence method. This model is presented as an example of the predictive capabilities of these kind of data-driven modelling techniques and to show how such models could be applied to karst in Ontario. The model was able to classify 70% of the validation dataset correctly while minimizing false positive identifications. This is moderately successful and could stand to be improved. Finally, suggestions to improving the current karst model of southern Ontario are suggested with the goal of increasing investigation into karst in Ontario and streamlining the reporting system for sinkholes, caves, and other karst features so as to improve the current Ontario karst database.

Spatio-temporal coverage optimization of sensor networks

Les réseaux de capteurs sont formés d’un ensemble de dispositifs capables de prendre individuellement des mesures d’un environnement particulier et d’échanger de l’information afin d’obtenir une représentation de haut niveau sur les activités en cours dans la zone d’intérêt. Une telle détection distribuée, avec de nombreux appareils situés à proximité des phénomènes d’intérêt, est pertinente dans des domaines tels que la surveillance, l’agriculture, l’observation environnementale, la surveillance industrielle, etc. Nous proposons dans cette thèse plusieurs approches pour effectuer l’optimisation des opérations spatio-temporelles de ces dispositifs, en déterminant où les placer dans l’environnement et comment les contrôler au fil du temps afin de détecter les cibles mobiles d’intérêt. La première nouveauté consiste en un modèle de détection réaliste représentant la couverture d’un réseau de capteurs dans son environnement. Nous proposons pour cela un modèle 3D probabiliste de la capacité de détection d’un capteur sur ses abords. Ce modèle inègre également de l’information sur l’environnement grâce à l’évaluation de la visibilité selon le champ de vision. À partir de ce modèle de détection, l’optimisation spatiale est effectuée par la recherche du meilleur emplacement et l’orientation de chaque capteur du réseau. Pour ce faire, nous proposons un nouvel algorithme basé sur la descente du gradient qui a été favorablement comparée avec d’autres méthodes génériques d’optimisation «boites noires» sous l’aspect de la couverture du terrain, tout en étant plus efficace en terme de calculs. Une fois que les capteurs placés dans l’environnement, l’optimisation temporelle consiste à bien couvrir un groupe de cibles mobiles dans l’environnement. D’abord, on effectue la prédiction de la position future des cibles mobiles détectées par les capteurs. La prédiction se fait soit à l’aide de l’historique des autres cibles qui ont traversé le même environnement (prédiction à long terme), ou seulement en utilisant les déplacements précédents de la même cible (prédiction à court terme). Nous proposons de nouveaux algorithmes dans chaque catégorie qui performent mieux ou produits des résultats comparables par rapport aux méthodes existantes. Une fois que les futurs emplacements de cibles sont prédits, les paramètres des capteurs sont optimisés afin que les cibles soient correctement couvertes pendant un certain temps, selon les prédictions. À cet effet, nous proposons une méthode heuristique pour faire un contrôle de capteurs, qui se base sur les prévisions probabilistes de trajectoire des cibles et également sur la couverture probabiliste des capteurs des cibles. Et pour terminer, les méthodes d’optimisation spatiales et temporelles proposées ont été intégrées et appliquées avec succès, ce qui démontre une approche complète et efficace pour l’optimisation spatio-temporelle des réseaux de capteurs.

Flexible modeling and estimation for high-dimensional graphs

Thesis (Ph.D.)--University of Washington, 2016-08

Learning Binary Code Representations for Effective and Efficient Image Retrieval

The size of online image datasets is constantly increasing. Considering an image dataset with millions of images, image retrieval becomes a seemingly intractable problem for exhaustive similarity search algorithms. Hashing methods, which encodes high-dimensional descriptors into compact binary strings, have become very popular because of their high efficiency in search and storage capacity. In the first part, we propose a multimodal retrieval method based on latent feature models. The procedure consists of a nonparametric Bayesian framework for learning underlying semantically meaningful abstract features in a multimodal dataset, a probabilistic retrieval model that allows cross-modal queries and an extension model for relevance feedback. In the second part, we focus on supervised hashing with kernels. We describe a flexible hashing procedure that treats binary codes and pairwise semantic similarity as latent and observed variables, respectively, in a probabilistic model based on Gaussian processes for binary classification. We present a scalable inference algorithm with the sparse pseudo-input Gaussian process (SPGP) model and distributed computing. In the last part, we define an incremental hashing strategy for dynamic databases where new images are added to the databases frequently. The method is based on a two-stage classification framework using binary and multi-class SVMs. The proposed method also enforces balance in binary codes by an imbalance penalty to obtain higher quality binary codes. We learn hash functions by an efficient algorithm where the NP-hard problem of finding optimal binary codes is solved via cyclic coordinate descent and SVMs are trained in a parallelized incremental manner. For modifications like adding images from an unseen class, we propose an incremental procedure for effective and efficient updates to the previous hash functions. Experiments on three large-scale image datasets demonstrate that the incremental strategy is capable of efficiently updating hash functions to the same retrieval performance as hashing from scratch.

Probability and Common-Sense: Tandem Towards Robust Robotic Object Recognition in Ambient Assisted Living

The suitable operation of mobile robots when providing Ambient Assisted Living (AAL) services calls for robust object recognition capabilities. Probabilistic Graphical Models (PGMs) have become the de-facto choice in recognition systems aiming to e ciently exploit contextual relations among objects, also dealing with the uncertainty inherent to the robot workspace. However, these models can perform in an inco herent way when operating in a long-term fashion out of the laboratory, e.g. while recognizing objects in peculiar con gurations or belonging to new types. In this work we propose a recognition system that resorts to PGMs and common-sense knowledge, represented in the form of an ontology, to detect those inconsistencies and learn from them. The utilization of the ontology carries additional advantages, e.g. the possibility to verbalize the robot's knowledge. A primary demonstration of the system capabilities has been carried out with very promising results.

Coordinated probabilistic optimal decision-making model for multi-area ATC with risk control

In the decision-making of multi-area ATC (Available Transfer Capacity) in electricity market environment, the existing resources of transmission network should be optimally dispatched and coordinately employed on the premise that the secure system operation is maintained and risk associated is controllable. The non-sequential Monte Carlo simulation is used to determine the ATC probability density distribution of specified areas under the influence of several uncertainty factors, based on which, a coordinated probabilistic optimal decision-making model with the maximal risk benefit as its objective is developed for multi-area ATC. The NSGA-II is applied to calculate the ATC of each area, which considers the risk cost caused by relevant uncertainty factors and the synchronous coordination among areas. The essential characteristics of the developed model and the employed algorithm are illustrated by the example of IEEE 118-bus test system. Simulative result shows that, the risk of multi-area ATC decision-making is influenced by the uncertainties in power system operation and the relative importance degrees of different areas.

A two-stage item recommendation method using probabilistic ranking with reconstructed tensor model

In a tag-based recommender system, the multi-dimensional correlation should be modeled effectively for finding quality recommendations. Recently, few researchers have used tensor models in recommendation to represent and analyze latent relationships inherent in multi-dimensions data. A common approach is to build the tensor model, decompose it and, then, directly use the reconstructed tensor to generate the recommendation based on the maximum values of tensor elements. In order to improve the accuracy and scalability, we propose an implementation of the -mode block-striped (matrix) product for scalable tensor reconstruction and probabilistically ranking the candidate items generated from the reconstructed tensor. With testing on real-world datasets, we demonstrate that the proposed method outperforms the benchmarking methods in terms of recommendation accuracy and scalability.

Probabilistic extension to realistic abductive reasoning model

In this paper, we give a method for probabilistic assignment to the Realistic Abductive Reasoning Model, The knowledge is assumed to be represented in the form of causal chaining, namely, hyper-bipartite network. Hyper-bipartite network is the most generalized form of knowledge representation for which, so far, there has been no way of assigning probability to the explanations, First, the inference mechanism using realistic abductive reasoning model is briefly described and then probability is assigned to each of the explanations so as to pick up the explanations in the decreasing order of plausibility.