246 research outputs found
Le pool des agences d'information des pays non-alignés
No full textObradovic (Обрадовић) Slobodan (Слободан). Le pool des agences d'information des pays non-alignés. In: Communication Information, volume 3 n°2, hiver 1980. L'information internationale : commerce ou propagande ? pp. 71-83
Forming the Cupolae With Concave Polyhedral Surfaces by Corrugating a Fourfold Strip of Equilateral Triangles
No full textThe cupolae with concave polyhedral surfaces consist of: two regular
polygons, n-gone and 2n-gone in parallel planes, interconnected by an
envelope constituted of series of equilateral triangles. The paper
describes cupolae which originate by corrugating of a fourfold strip of
equilateral triangles, forming thereby the envelope of a cupola. In
this manner, a non-convex polyhedron is emerged. Such a method of corrugating the envelope, allows the solutions for generating cupolae with base polygon which number of sides exceedes n=10, which was the maximal number of sides for cupole with the envelope consisted of twofold strip of equilateral triangles. By analyzing the elements of these polyhedra and by help of their paper models, we find geometric constructions and projection procedures by which it is possible to graphically display the cupolae.http://mongeometrija.com/media/mongeometrija/2010/moNGeometrija%202010%20-%20PAGINACIJA.pdf
http://mongeometrija.com/component/content/article/52-slobodan-mii/115-misic-obradovic-forming-the-cupolae-with-concave-polyhedral-surfaces-by-corrugating-a-fourfold-s
Coping with Missing and Incomplete Information in Natural Language Processing with Applications in Sentiment Analysis and Entity Matching
No full textMuch work in Natural Language Processing (NLP) is broadly concerned with extracting useful information from unstructured text passages. In recent years there has been an increased focus on informal writing as is found in online venues such as Twitter and Yelp. Processing this text introduces additional difficulties for NLP techniques, for example, many of the terms may be unknown due to rapidly changing vocabulary usage. A straightforward NLP approach will not have any capability of using the information these terms provide. In such \emph{information poor} environments of missing and incomplete information, it is necessary to develop novel, clever methods for leveraging the information we have explicitly available to unlock key nuggets of implicitly available information. In this work we explore several such methods and how they can collectively help to improve NLP techniques in general, with a focus on Sentiment Analysis (SA) and Entity Matching (EM). The problem of SA is that of identifying the polarity (positive, negative, neutral) of a speaker or author towards the topic of a given piece of text. SA can focus on various levels of granularity. These include finding the overall sentiment of a long text document, finding the sentiment of individual sentences or phrases, or finding the sentiment directed toward specific entities and their aspects (attributes). The problem of EM, also known as Record Linkage, is the problem of determining records from independent and uncooperative data sources that refer to the same real-world entities. Traditional approaches to EM have used the record representation of entities to accomplish this task. With the nascence of social media, entities on the Web are now accompanied by user generated content, which allows us to apply NLP solutions to the problem. We investigate specifically the following aspects of NLP for missing and incomplete information: (1) Inferring a sentiment polarity (i.e., the positive, negative, and neutral composition) of new terms. (2) Inferring a representation of new vocabulary terms that allows us to compare these terms with known terms in regards to their meaning and sentiment orientation. This idea can be further expanded to derive the representation of larger chunks of text, such as multi-word phrases. (3) Identifying key attributes of highly salient sentiment bearing passages that allow us to identify such sections of a document, even when the complete text is not analyzable. (4) Using text based methods to match corresponding entities (e.g., restaurants or hotels) from independent data sources that may miss key identifying attributes such as names or addresses.Computer and Information Scienc
NOVEL DATA MINING ALGORITHMS FOR ANALYSIS OF ELECTRONIC HEALTH RECORDS
No full textMedical health providers use electronic health records (EHRs) to store information about patient treatment to support patient care management and securely share health information among healthcare organizations. EHRs have also been used in healthcare research in problems such as patient phenotyping, health risk prediction, and medical entity extraction. In this thesis, we focus on several important issues: (1) how to convert natural text from medical notes to vector representations suitable for deep learning algorithms, (2) how to help healthcare researchers select a patient cohort from EHRs, and (3) how to use EHRs to identify patient diagnoses and treatments.
In the first part of the thesis, we present a new method for learning vector representations of medical terms. Learning vector representations of words is an important pre-processing step in many natural language processing applications. For example, EHRs contain clinical notes that describe patient health conditions and course of treatment in a narrative style. The notes contain specialized medical terminology and many abbreviations. Learning good vector representations of specialized medical terms can improve the quality of downstream data analysis tasks on EHR data. However, the traditional approaches struggle to learn vector representations of rarely used medical terms. To overcome this problem, we developed a neural network-based approach, called definition2vec, that uses external knowledge contained in medical vocabularies. We performed quantitative and qualitative analysis to measure the usefulness of the learned representations. The results demonstrate that definition2vec is superior to the state-of-the-art algorithms.
In the second part of the thesis, we describe a new visual interface that helps healthcare researchers select patient cohorts from EHR data. Process of identifying patients of interest for observational studies from EHR data is known as cohort selection, a challenging research problem. We considered a problem of cohort selection from medical claim data, which requires identifying a set of medical codes for selection. However, there are tens of thousands of unique medical codes, and it becomes very difficult for any human to decide which codes identify patients of interest. To help users in defining a set of codes for cohort identification, we developed an interactive system, called Medical Claim Visualization system (MedCV), which visualizes medical code representations. MedCV analyzes a medical claim database and allows users to reason about medical code relationships and define inclusion rules for the selection by visualizing medical codes, claims, and patient timelines. Evaluation of our system through a user study indicates that MedCV enables domain experts to define inclusion rules efficiently and with high quality.
The third part of the thesis is a study of the definition of acute kidney injury (AKI), which is a condition where kidneys suddenly cannot filter waste from the blood. AKI is a major cause of patient death in intensive care units (ICU) and it is critical to detect it early. Recently published KDIGO medical guideline proposed a clinical definition of AKI using blood serum creatinine and urine output. The KDIGO definition was developed based on the expert knowledge, but very little is known about how well it matches the medical practice. In this study, we investigated publicly available EHR data from 47,499 ICU admissions to determine the concordance between the KDIGO definition and AKI determination by the medical provider. We show that it is possible to find a formula using machine learning with much higher concordance with the medical provider AKI coding than KDIGO and discuss the medical relevance of this finding.Computer and Information Scienc
Quality control of the hindside's axis stabilization using the instrumental servo system
No full textIn up to date wars, gunfiring from the vehicle in motion has become the basic way of action for operational armored vehicles (conveyors and tanks). Because of the rolling of the vehicle's body (caused by moving on the ground) and the change in the target coordinates (because the target is in motion, as well), the shooting accuracy in comparison with shooting accuracy in the case of shooting passive targets from the fixed position is substantially decreased. The particular problem is to measure the quality that is, to measure the error of achieved stabilization. This study describes the new, simple (and cheap) simulation method for the stabilization quality check, which enables the complete laboratory simulation of driving over the optional terrain in the real conditions, as well as the undisturbed examining in the extreme temperature conditions.</jats:p
GAUSSIAN CONDITIONAL RANDOM FIELDS FOR REGRESSION IN REMOTE SENSING
No full textIn recent years many remote sensing instruments of various properties have been employed in an attempt to better characterize important geophysical phenomena. Satellite instruments provide an exceptional opportunity for global long-term observations of the land, the biosphere, the atmosphere, and the oceans. The collected data are used for estimation and better understanding of geophysical parameters such as land cover type, atmospheric properties, or ocean temperature. Achieving accurate estimations of such parameters is an important requirement for development of models able to predict global climate changes. One of the most challenging climate research problems is estimation of global composition, load, and variability of aerosols, small airborne particles that reflect and absorb incoming solar radiation. The existing algorithm for aerosol prediction from satellite observations is deterministic and manually tuned by domain scientist. In contrast to domain-driven method, we show that aerosol prediction is achievable by completely data-driven approaches. These statistical methods consist of learning of nonlinear regression models to predict aerosol load using the satellite observations as inputs. Measurements from unevenly distributed ground-based sites over the world are used as proxy to ground-truth outputs. Although statistical methods achieve better accuracy than deterministic method this setup is appropriate when data are independently and identically distributed (IID). The IID assumption is often violated in remote sensing where data exhibit temporal, spatial, or spatio-temporal dependencies. In such cases, the traditional supervised learning approaches could result in a model with degraded accuracy. Conditional random fields (CRF) are widely used for predicting output variables that have some internal structure. Most of the CRF research has been done on structured classification where the outputs are discrete. We propose a CRF model for continuous outputs that uses multiple unstructured predictors to form its features and at the same time exploits structure among outputs. By constraining the feature functions to quadratic functions of outputs, we show that the CRF model can be conveniently represented in a Gaussian canonical form. The appeal of proposed Gaussian Conditional Random Fields (GCRF) model is in its conceptual simplicity and computational efficiency of learning and inference through use of sparse matrix computations. Experimental results provide strong evidence that the GCRF achieves better accuracy than non-structured models. We improve the representational power of the GCRF model by 1) introducing the adaptive feature function that can learn nonlinear relationships between inputs and outputs and 2) allowing the weights of feature functions to be dependent on inputs. The GCRF is also readily applicable to other regression applications where there is a need for knowledge integration, data fusion, and exploitation of correlation among output variables.Computer and Information Scienc
Bayesian Sparse Regression with Application to Data-driven Understanding of Climate
No full textSparse regressions based on constraining the L1-norm of the coefficients became popular due to their ability to handle high dimensional data unlike the regular regressions which suffer from overfitting and model identifiability issues especially when sample size is small. They are often the method of choice in many fields of science and engineering for simultaneously selecting covariates and fitting parsimonious linear models that are better generalizable and easily interpretable. However, significant challenges may be posed by the need to accommodate extremes and other domain constraints such as dynamical relations among variables, spatial and temporal constraints, need to provide uncertainty estimates and feature correlations, among others. We adopted a hierarchical Bayesian version of the sparse regression framework and exploited its inherent flexibility to accommodate the constraints. We applied sparse regression for the feature selection problem of statistical downscaling of the climate variables with particular focus on their extremes. This is important for many impact studies where the climate change information is required at a spatial scale much finer than that provided by the global or regional climate models. Characterizing the dependence of extremes on covariates can help in identification of plausible causal drivers and inform extremes downscaling. We propose a general-purpose sparse Bayesian framework for covariate discovery that accommodates the non-Gaussian distribution of extremes within a hierarchical Bayesian sparse regression model. We obtain posteriors over regression coefficients, which indicate dependence of extremes on the corresponding covariates and provide uncertainty estimates, using a variational Bayes approximation. The method is applied for selecting informative atmospheric covariates at multiple spatial scales as well as indices of large scale circulation and global warming related to frequency of precipitation extremes over continental United States. Our results confirm the dependence relations that may be expected from known precipitation physics and generates novel insights which can inform physical understanding. We plan to extend our model to discover covariates for extreme intensity in future. We further extend our framework to handle the dynamic relationship among the climate variables using a nonparametric Bayesian mixture of sparse regression models based on Dirichlet Process (DP). The extended model can achieve simultaneous clustering and discovery of covariates within each cluster. Moreover, the a priori knowledge about association between pairs of data-points is incorporated in the model through must-link constraints on a Markov Random Field (MRF) prior. A scalable and efficient variational Bayes approach is developed to infer posteriors on regression coefficients and cluster variables.Computer and Information Scienc
Learning from Structured Data: Scalability, Stability and Temporal Awareness
No full textA plethora of high-impact applications involve predictive modeling of structured data. In various domains, from hospital readmission prediction in the medical realm, though weather forecasting and event detection in power systems, up to conversion prediction in online businesses, the data holds a certain underlying structure. Building predictive models from such data calls for leveraging the structure as an additional source of information. Thus, a broad range of structure-aware approaches have been introduced, yet certain common challenges in many structured learning scenarios remain unresolved. This dissertation revolves around addressing the challenges of scalability, algorithmic stability and temporal awareness in several scenarios of learning from either graphically or sequentially structured data.
Initially, the first two challenges are discussed from a structured regression standpoint. The studies addressing these challenges aim at designing scalable and algorithmically stable models for structured data, without compromising their prediction performance. It is further inspected whether such models can be applied to both static and dynamic (time-varying) graph data. To that end, a structured ensemble model is proposed to scale with the size of temporal graphs, while making stable and reliable yet accurate predictions on a real-world application involving gene expression prediction. In the case of static graphs, a theoretical insight is provided on the relationship between algorithmic stability and generalization in a structured regression setting. A stability-based objective function is designed to indirectly control the stability of a collaborative ensemble regressor, yielding generalization performance improvements on structured regression applications as diverse as predicting housing prices based on real-estate transactions and readmission prediction from hospital records.
Modeling data that holds a sequential rather than a graphical structure requires addressing temporal awareness as one of the major challenges. In that regard, a model is proposed to generate time-aware representations of user activity sequences, intended to be seamlessly applicable across different user-related tasks, while sidestepping the burden of task-driven feature engineering. The quality and effectiveness of the time-aware user representations led to predictive performance improvements over state-of-the-art models on multiple large-scale conversion prediction tasks.
Sequential data is also analyzed from the perspective of a high-impact application in the realm of power systems. Namely, detecting and further classifying disturbance events, as an important aspect of risk mitigation in power systems, is typically centered on the challenges of capturing structural characteristics in sequential synchrophasor recordings. Therefore, a thorough comparative analysis was conducted by assessing various traditional as well as more sophisticated event classification models under different domain-expert-assisted labeling scenarios. The experimental findings provide evidence that hierarchical convolutional neural networks (HCNNs), capable of automatically learning time-invariant feature transformations that preserve the structural characteristics of the synchrophasor signals, consistently outperform traditional model variants. Their performance is observed to further improve as more data are inspected by a domain expert, while smaller fractions of solely expert-inspected signals are already sufficient for HCNNs to achieve satisfactory event classification accuracy. Finally, insights into the impact of the domain expertise on the downstream classification performance are also discussed.Computer and Information Scienc
Learning from multi-modal spatiotemporal data: machine learning approaches to advance resilience in smart grids
No full textThe electric grid has been expanding both in size and the technologies used. As of the 2020s, the United States power grid consists of more than 9,200 electric generating units with more than 1 million megawatts of generating capacity connected to more than 300,000 miles of transmission lines. The United States electricity grid has rapidly expanded in recent decades, and the majority (over 70\%) of its infrastructure has exceeded 25 years of age. Due to its size and age, several challenges have emerged. Widespread power outages have been increasing across the United States. Between 2018 and 2020, more than 231,000 power outages occurred in the United States that lasted more than one hour, out of which 17,484 lasted at least eight hours. In the same period, the power outages resulted in an annual loss of 520 million customer hours across 2,447 U.S. counties. Moreover, and with the rapidly changing climate, between 2000 and 2021, approximately 83\% of significant power outages impacting a minimum of 50,000 customers in the United States were attributed to severe weather conditions. Lastly, the increasing use of renewables and other non-traditional generation methods forces the power system towards a more decentralized model, with many integrated systems constantly added to the grid. This decentralization adds additional burdens on controlling systems and grid operators. The rapid growth of technology and data storage allowed the deployment of sensing devices across the electric grid. Such technologies present a golden opportunity to tackle many of the electric grid's challenges. Despite that, such technologies presented many challenges simultaneously. With the large amounts of data, it became humanly impossible to comprehend, analyze, and use all collected data manually. While machine learning can be used to analyze smart grid data, this can be challenged by the nature of its data. Smart grid produces high-dimensional spatiotemporal data, and many applications require multi-modal data. Moreover, power systems' data quality challenges add complexities to model development. The data is noisy, contains missing segments, and usually has incomplete and inaccurate labels. In addition, interpreting machine learning models in the context of smart grids poses unique challenges. To address these challenges, different models for multiple smart-grid applications were introduced in this research, where each model focused on producing practical solutions for the challenges facing current-day smart grids. Using spatiotemporal data, a solar generation prediction model was proposed. The solution combined spatial and temporal data, then utilized machine learning embeddings to build datasets to train downstream models. This resulted in accurate prediction of solar generation across several settings. In addition to solar generation prediction, several models were introduced to detect, predict and explain power grid faults. A neural model is introduced to detect power faults from Phasor Measurement Unit (PMU) data. A novel method is introduced to preprocesses, de-noise, and combine high dimensional data, then this data is used to train novel neural methods that detect faults in multiple settings. This model addressed issues of high dimensionality and data quality. After that, several models studying power fault prediction and precursor discovery were introduced. A model that jointly predicts outages 6 hours ahead and produces explainable event precursors from multi-modal data is introduced. Where such precursors can assist power grid operators to take action to mitigate widespread power outages. Finally, a novel methodology is introduced that expands to previous work by predicting and extracting event precursors spatiotemporally 12 hours in advance. Where event precursors can be predicted on multiple spatial locations simultaneously, extracted spatiotemporal event precursors can help grid operators narrow down mitigation plans and help reduce the risk of widespread power outages.Computer and Information Scienc
LEVERAGING TEMPORAL SUBSEQUENCES FOR TIME-SERIES CLASSIFICATION
No full textResearch on time-series classification has garnered importance among practitioners in the data mining community. A major reason behind the ever-increasing interest among data-miners is the plethora of time-series data available from a wide range of real-life domains. Temporal-ordered data from a variety of sensor-based domains such as wearable devices, smart homes, industrial monitoring, medical diagnosis, etc. provide classification challenges more akin to real-world scenarios. Thus, building more robust time-series classification models is imperative. One group of popular models focuses on identifying short discriminative temporal patterns (subsequences) from the time-series for classification. These temporal subsequences, known as shapelets, are local patterns that can be used to uniquely identify the target class of a time-series instance. In this dissertation, I explore two real-world challenges pertaining to shapelet based time-series classification models and provide solutions to mitigate those challenges. In the first challenge, the problem of cost-sensitive learning in time-series classification is explored. First, the problem of highly imbalanced time-series classification using shapelets is investigated. The current state-of-the-art approach learns generalized shapelets along with weights of the classification hyperplane via a classical cost-insensitive loss function. Cost-insensitive loss functions tend to treat different misclassification errors equally, and thus, models are usually biased towards examples of the majority class. In this research, the generalized shapelets learning framework is extended and a cost-sensitive learning model is proposed. Instead of incorporating the misclassification cost as prior knowledge, as was done by other published methods, a constrained optimization problem was formulated to learn the unknown misclassification costs along with the shapelets and their weights. Secondly, I focus on the problem of cost-sensitive early classification in time-series datasets. High false alarm rates in intensive care units (ICUs) cause desensitization among care providers, thus risking patients' lives. Providing early detection of true and false cardiac arrhythmia alarms can alert hospital personnel and avoid alarm fatigue. This will ensure hospital personnel can act only on true life-threatening alarms, hence improving efficiency in ICUs. Furthermore, suppressing false alarms cannot be an excuse to suppress true alarm detection rates. In this study, a cost-sensitive approach for false alarm suppression while keeping near perfect true alarm detection rates was investigated using a confidence estimate for shapelets matching. In the second challenge, the temporal dependencies among shapelets are explored. The existing shapelet-based methods for time-series classification assume that shapelets are independent of each other. However, they neglect temporal dependencies among pairs of shapelets, which are informative features that exist in many applications. Within this new framework, a scheme is explored to extract informative orders among shapelets by considering the time gap between pairs of shapelets. In this realm, two models are proposed, Pairwise Shapelet-Orders Discovery (PSOD) and Learning pairwise Orders and Shapelets (LOS), which extracts both informative shapelets and shapelet-orders and incorporates the shapelet-transformed space with shapelet-order space for time-series classification. The two proposed models are contrasting approaches in the time-series classification paradigm. The PSOD is a search-based greedy procedure to extract unique shapelets and identify orders among the selected shapelets. On the other hand, LOS is an optimization-based approach to extract shapelet-orders among learned generalized shapelets. However, in both the hypotheses, the extracted pairwise shapelet-orders could increase the confidence of the prediction and further improve the classification performance. The experimental results provide evidence that when considering shapelet-orders, classification accuracy is significantly improved on average over baseline methods. To the best of my knowledge, these are the first work that proposes formal methodologies to extract shapelet-orders and present augmented space of shapelets and shapelet-orders.Computer and Information Scienc
- …
