Obwohl schon seit den 50er Jahren auf dem Gebiet der Automatischen Textzusammenfassung Forschung betrieben wird, wurden der Nutzen und die Notwendigkeit dieser Systeme erst mit dem Boom des Internets richtig erkannt. Das World Wide Web stellt eine täglich wachsende Menge an Informationen zu nahezu jedem Thema zur Verfügung. Um den Zeitaufwand zum Finden und auch zum Wiederfinden der richtigen Informationen zu minimieren, traten Suchmaschinen ihren Siegeszug an. Doch um einen Überblick zu einem ausgewählten Thema zu erhalten, ist eine einfache Auflistung aller in Frage kommenden Seiten nicht mehr adäquat. Zusätzliche Mechanismen wie Extraktionsalgorithmen für die automatische Generierung von Zusammenfassungen können hier helfen, Suchmaschinen oder Webkataloge zu optimieren, um so den Zeitaufwand bei der Recherche zu verringern und die Suche einfacher und komfortabler zu gestalten. In dieser Diplomarbeit wurde eine Analyse von Extraktionsalgorithmen durchgeführt, welche für die automatische Textzusammenfassung genutzt werden können. Auf Basis dieser Analyse als viel versprechend eingestufte Algorithmen wurden in Java implementiert und die mit diesen Algorithmen erstellten Zusammenfassungen in einer Evaluation verglichen.

Thesis (Ph. D.)--Johns Hopkins University. / Vita. Includes bibliographical references.

Zugl.: Klagenfurt, Univ., Diss., 2009

Communities have traditionally been viewed as spatially bounded social units composed of multiple households whose inhabitants are integrated by shared resources and a common sense of identity. While investigating resources and identity is useful for archaeological study because of their material correlates, such views of community ultimately fail to acknowledge the dynamic interaction between cultural and environmental forces in shaping and shifting those arrangements over time. This study examines settlement, excavation, and geoarchaeological data from the Palmarejo Valley in northwestern Honduras with the aim of modeling the process of community formation at the intersection of social and natural landscapes in both the past and present.

A primary goal in systems neuroscience is to understand how neural spike responses encode information about the external world. A popular approach to this problem is to build an explicit probabilistic model that characterizes the encoding relationship in terms of a cascade of stages: (1) linear dimensionality reduction of a high-dimensional stimulus space using a bank of filters or receptive fields (RFs); (2) a nonlinear function from filter outputs to spike rate; and (3) a stochastic spiking process with recurrent feedback. These models have described single- and multi-neuron spike responses in a wide variety of brain areas. This dissertation addresses Bayesian methods to efficiently estimate the linear and non-linear stages of the cascade encoding model. In the first part, the dissertation describes a novel Bayesian receptive field estimator based on a hierarchical prior that flexibly incorporates knowledge about the shapes of neural receptive fields. This estimator achieves error rates several times lower than existing methods, and can be applied to a variety of other neural inference problems such as extracting structure in fMRI data. The dissertation also presents active learning frameworks developed for receptive field estimation incorporating a hierarchical prior in real-time neurophysiology experiments. In addition, the dissertation describes a novel low-rank model for the high dimensional receptive field, combined with a hierarchical prior for more efficient receptive field estimation. In the second part, the dissertation describes new models for neural nonlinearities using Gaussian processes (GPs) and Bayesian active learning algorithms in closed-loop neurophysiology experiments to rapidly estimate neural nonlinearities. The dissertation also presents several stimulus selection criteria and compare their performance in neural nonlinearity estimation. Furthermore, the dissertation presents a variation of the new models by including an additional latent Gaussian noise source, to infer the degree of over-dispersion in neural spike responses. The proposed model successfully captures various mean-variance relationships in neural spike responses and achieves higher prediction accuracy than previous models. / text

Localization is critical to a number of wireless network applications. In many situations GPS is not suitable. This dissertation (i) develops novel localization schemes for wireless networks by explicitly incorporating mobility information and (ii) applies localization to physical analytics i.e., understanding shoppers' behavior within retail spaces by leveraging inertial sensors, Wi-Fi and vision enabled by smart glasses. More specifically, we first focus on multi-hop mobile networks, analyze real mobility traces and observe that they exhibit temporal stability and low-rank structure. Motivated by these observations, we develop novel localization algorithms to effectively capture and also adapt to different degrees of these properties. Using extensive simulations and testbed experiments, we demonstrate the accuracy and robustness of our new schemes. Second, we focus on localizing a single mobile node, which may not be connected with multiple nodes (e.g., without network connectivity or only connected with an access point). We propose trajectory-based localization using Wi-Fi or magnetic field measurements. We show that these measurements have the potential to uniquely identify a trajectory. We then develop a novel approach that leverages multi-level wavelet coefficients to first identify the trajectory and then localize to a point on the trajectory. We show that this approach is highly accurate and power efficient using indoor and outdoor experiments. Finally, localization is a critical step in enabling a lot of applications --- an important one is physical analytics. Physical analytics has the potential to provide deep-insight into shoppers' interests and activities and therefore better advertisements, recommendations and a better shopping experience. To enable physical analytics, we build ThirdEye system which first achieves zero-effort localization by leveraging emergent devices like the Google-Glass to build AutoLayout that fuses video, Wi-Fi, and inertial sensor data, to simultaneously localize the shoppers while also constructing and updating the product layout in a virtual coordinate space. Further, ThirdEye comprises of a range of schemes that use a combination of vision and inertial sensing to study mobile users' behavior while shopping, namely: walking, dwelling, gazing and reaching-out. We show the effectiveness of ThirdEye through an evaluation in two large retail stores in the United States. / text

Many problems in signal processing, machine learning and computer vision can be solved by learning low rank models from data. In computer vision, problems such as rigid structure from motion have been formulated as an optimization over subspaces with fixed rank. These hard-rank constraints have traditionally been imposed by a factorization that parameterizes subspaces as a product of two matrices of fixed rank. Whilst factorization approaches lead to efficient and kernelizable optimization algorithms, they have been shown to be NP-Hard in presence of missing data. Inspired by recent work in compressed sensing, hard-rank constraints have been replaced by soft-rank constraints, such as the nuclear norm regularizer. Vis-a-vis hard-rank approaches, soft-rank models are convex even in presence of missing data: but how is convex optimization solving a NP-Hard problem? This thesis addresses this question by analyzing the relationship between hard and soft rank constraints in the unsupervised factorization with missing data problem. Moreover, we extend soft rank models to weakly supervised and fully supervised learning problems in computer vision. There are four main contributions of our work: (1) The analysis of a new unified low-rank model for matrix factorization with missing data. Our model subsumes soft and hard-rank approaches and merges advantages from previous formulations, such as efficient algorithms and kernelization. It also provides justifications on the choice of algorithms and regions that guarantee convergence to global minima. (2) A deterministic \rank continuation" strategy for the NP-hard unsupervised factorization with missing data problem, that is highly competitive with the state-of-the-art and often achieves globally optimal solutions. In preliminary work, we show that this optimization strategy is applicable to other NP-hard problems which are typically relaxed to convex semidentite programs (e.g., MAX-CUT, quadratic assignment problem). (3) A new soft-rank fully supervised robust regression model. This convex model is able to deal with noise, outliers and missing data in the input variables. (4) A new soft-rank model for weakly supervised image classification and localization. Unlike existing multiple-instance approaches for this problem, our model is convex.

This thesis explores extensions of Random Utility Models (RUMs), providing more flexible models and adopting a computational perspective. This includes building new models and understanding their properties such as identifiability and the log concavity of their likelihood functions as well as the development of estimation algorithms. / Engineering and Applied Sciences

Localization is a fundamental operation for many wireless networks. While GPS is widely used for location determination, it is unavailable in many environments either due to its high cost or the lack of line of sight to the satellites (e.g., indoors, under the ground, or in a downtown canyon). The limitations of GPS have motivated researchers to develop many localization schemes to infer locations based on measured wireless signals. However, most of these existing schemes focus on localization in static wireless networks. As many wireless networks are mobile (e.g., mobile sensor networks, disaster recovery networks, and vehicular networks), we focus on localization in mobile networks in this thesis. We analyze real mobility traces and find that they exhibit temporal stability and low-rank structure. Motivated by this observation, we develop three novel localization schemes to accurately determine locations in mobile networks: 1. Low Rank based Localization (LRL), which exploits the low-rank structure in mobility. 2. Temporal Stability based Localization (TSL), which leverages the temporal stability. 3. Temporal Stability and Low Rank based Localization (TSLRL), which incorporates both the temporal stability and the low-rank structure. These localization schemes are general and can leverage either mere connectivity (i.e., range-free localization) or distance estimation between neighbors (i.e., range-based localization). Using extensive simulations and testbed experiments, we show that our new schemes significantly outperform state-of-the-art localization schemes under a wide range of scenarios and are robust to measurement errors. / text

If a structure is composed mainly of one material but contains a small amount of a second material, and if these two materials have significantly different levels of structural damping, this can increase the cost of solving the modal frequency response problem substantially. Even if the rank of the contribution to the finite element structural damping matrix from the second material is very low, the matrix becomes fully populated when transformed to the modal representation. As a result, the complex-valued modal matrix that represents the structure’s stiffness and structural damping is both full rank, because of the diagonal part contributed by the stiffness, and fully populated, because of off-diagonal imaginary terms contributed by the second material’s structural damping. Solving the modal frequency response problem at many frequencies requires either the factorization of a coefficient matrix at every frequency, or the solution of a complex symmetric eigenvalue problem associated with the modal stiffness/structural damping matrix. The cost of both of these approaches is proportional to the cube of the number of modes included in the analysis. This cost could be reduced greatly if the damping properties of the structure were handled carefully in modeling the structure, but in practical computation of the modal frequency response, the information that could potentially reduce the computational cost is often unavailable. This thesis explores the possibilities of obtaining a representation of the complex modal stiffness/structural damping matrix as a diagonal matrix plus a matrix of minimal rank. An algorithm for computing a “diagonal plus low rank” (DPLR) representation is developed, along with an iterative algorithm for using an inexact DPLR approximation in the solution of the modal frequency response problem. The behavior of these algorithms is investigated on several example problems. / text