ImageRover: A Content-Based Image Browser for the World Wide Web

ImageRover is a search by image content navigation tool for the world wide web. To gather images expediently, the image collection subsystem utilizes a distributed fleet of WWW robots running on different computers. The image robots gather information about the images they find, computing the appropriate image decompositions and indices, and store this extracted information in vector form for searches based on image content. At search time, users can iteratively guide the search through the selection of relevant examples. Search performance is made efficient through the use of an approximate, optimized k-d tree algorithm. The system employs a novel relevance feedback algorithm that selects the distance metrics appropriate for a particular query.

Advanced glycation end products (AGEs) co-localize with AGE receptors in the retinal vasculature of diabetic and of AGE-infused rats

Advanced glycation end products (AGEs), formed from the nonenzymatic glycation of proteins and lipids with reducing sugars, have been implicated in many diabetic complications; however, their role in diabetic retinopathy remains largely unknown. Recent studies suggest that the cellular actions of AGEs may be mediated by AGE-specific receptors (AGE-R). We have examined the immunolocalization of AGEs and AGE-R components R1 and R2 in the retinal vasculature at 2, 4, and 8 months after STZ-induced diabetes as well as in nondiabetic rats infused with AGE bovine serum albumin for 2 weeks. Using polyclonal or monoclonal anti-AGE antibodies and polyclonal antibodies to recombinant AGE-R1 and AGE-R2, immunoreactivity (IR) was examined in the complete retinal vascular tree after isolation by trypsin digestion. After 2, 4, and 8 months of diabetes, there was a gradual increase in AGE IR in basement membrane. At 8 months, pericytes, smooth muscle cells, and endothelial cells of the retinal vessels showed dense intracellular AGE IR. AGE epitopes stained most intensely within pericytes and smooth muscle cells but less in basement membrane of AGE-infused rats compared with the diabetic group. Retinas from normal or bovine-serum-albumin-infused rats were largely negative for AGE IR. AGE-R1 and -R2 co-localized strongly with AGEs of vascular endothelial cells, pericytes, and smooth muscle cells of either normal, diabetic, or AGE-infused rat retinas, and this distribution did not vary with each condition. The data indicate that AGEs accumulate as a function of diabetes duration first within the basement membrane and then intracellularly, co-localizing with cellular AGE-Rs. Significant AGE deposits appear within the pericytes after long-term diabetes or acute challenge with AGE infusion conditions associated with pericyte damage. Co-localization of AGEs and AGE-Rs in retinal cells points to possible interactions of pathogenic significance.

A Hierarchical Cache Coherent Protocol

As the number of processors in distributed-memory multiprocessors grows, efficiently supporting a shared-memory programming model becomes difficult. We have designed the Protocol for Hierarchical Directories (PHD) to allow shared-memory support for systems containing massive numbers of processors. PHD eliminates bandwidth problems by using a scalable network, decreases hot-spots by not relying on a single point to distribute blocks, and uses a scalable amount of space for its directories. PHD provides a shared-memory model by synthesizing a global shared memory from the local memories of processors. PHD supports sequentially consistent read, write, and test- and-set operations. This thesis also introduces a method of describing locality for hierarchical protocols and employs this method in the derivation of an abstract model of the protocol behavior. An embedded model, based on the work of Johnson[ISCA19], describes the protocol behavior when mapped to a k-ary n-cube. The thesis uses these two models to study the average height in the hierarchy that operations reach, the longest path messages travel, the number of messages that operations generate, the inter-transaction issue time, and the protocol overhead for different locality parameters, degrees of multithreading, and machine sizes. We determine that multithreading is only useful for approximately two to four threads; any additional interleaving does not decrease the overall latency. For small machines and high locality applications, this limitation is due mainly to the length of the running threads. For large machines with medium to low locality, this limitation is due mainly to the protocol overhead being too large. Our study using the embedded model shows that in situations where the run length between references to shared memory is at least an order of magnitude longer than the time to process a single state transition in the protocol, applications exhibit good performance. If separate controllers for processing protocol requests are included, the protocol scales to 32k processor machines as long as the application exhibits hierarchical locality: at least 22% of the global references must be able to be satisfied locally; at most 35% of the global references are allowed to reach the top level of the hierarchy.

Application for fault location in electrical power distribution systems

Fault location has been studied deeply for transmission lines due to its importance in power systems. Nowadays the problem of fault location on distribution systems is receiving special attention mainly because of the power quality regulations. In this context, this paper presents an application software developed in Matlabtrade that automatically calculates the location of a fault in a distribution power system, starting from voltages and currents measured at the line terminal and the model of the distribution power system data. The application is based on a N-ary tree structure, which is suitable to be used in this application due to the highly branched and the non- homogeneity nature of the distribution systems, and has been developed for single-phase, two-phase, two-phase-to-ground, and three-phase faults. The implemented application is tested by using fault data in a real electrical distribution power system

Warped extra dimensions: flavor, precision tests and Higgs physics

In this thesis, the phenomenology of the Randall-Sundrum setup is investigated. In this context models with and without an enlarged SU(2)_L x SU(2)_R x U(1)_X x P_{LR} gauge symmetry, which removes corrections to the T parameter and to the Z b_L \bar b_L coupling, are compared with each other. The Kaluza-Klein decomposition is formulated within the mass basis, which allows for a clear understanding of various model-specific features. A complete discussion of tree-level flavor-changing effects is presented. Exact expressions for five dimensional propagators are derived, including Yukawa interactions that mediate flavor-off-diagonal transitions. The symmetry that reduces the corrections to the left-handed Z b \bar b coupling is analyzed in detail. In the literature, Randall-Sundrum models have been used to address the measured anomaly in the t \bar t forward-backward asymmetry. However, it will be shown that this is not possible within a natural approach to flavor. The rare decays t \to cZ and t \to ch are investigated, where in particular the latter could be observed at the LHC. A calculation of \Gamma_{12}^{B_s} in the presence of new physics is presented. It is shown that the Randall-Sundrum setup allows for an improved agreement with measurements of A_{SL}^s, S_{\psi\phi}, and \Delta\Gamma_s. For the first time, a complete one-loop calculation of all relevant Higgs-boson production and decay channels in the custodial Randall-Sundrum setup is performed, revealing a sensitivity to large new-physics scales at the LHC.

Calcareous nannofossil stratigraphy, nannofossil events and carbon isotopic ratios of ODP Leg 183 sites

Cores from Sites 1135, 1136, and 1138 of Ocean Drilling Program Leg 183 to the Kerguelen Plateau (KP) provide the most complete Paleocene and Eocene sections yet recovered from the southern Indian Ocean. These nannofossil-foraminifer oozes and chalks provide an opportunity to study southern high-latitude biostratigraphic and paleoceanographic events, which is the primary subject of this paper. In addition, a stable isotope profile was established across the Cretaceous/Tertiary (K/T) boundary at Site 1138. An apparently complete K/T boundary was recovered at Site 1138 in terms of assemblage succession, isotopic signature, and reworking of older (Cretaceous) nannofossil taxa. There is a significant color change, a negative carbon isotope shift, and nannofossil turnover. The placement of the boundary based on these criteria, however, is not in agreement with the available shipboard paleomagnetic stratigraphy. We await shore-based paleomagnetic study to confirm or deny those preliminary results. The Paleocene nannofossil assemblage is, in general, characteristic of the high latitudes with abundant Chiasmolithus, Prinsius, and Toweius. Placed in context with other Southern Ocean sites, the biogeography of Hornibrookina indicates the presence of some type of water mass boundary over the KP during the earliest Paleocene. This boundary disappeared by the late Paleocene, however, when there was an influx of warm-water discoasters, sphenoliths, and fasciculiths. This not only indicates that during much of the late Paleocene water temperatures were relatively equable, but preliminary floral and stable isotope analyses also indicate that a relatively complete record of the late Paleocene Thermal Maximum event was recovered at Site 1135. It was only at the beginning of the middle Eocene that water temperatures began to decline and the nannofossil assemblage became dominated by cool-water species while discoaster and sphenolith abundances and diversity were dramatically reduced. One new taxonomic combination is proposed, Heliolithus robustus Arney, Ladner, and Wise.

Optimization and Machine Learning Frameworks for Complex Network Analysis

Thesis (Ph.D.)--University of Washington, 2016-06

On reliable computation by noisy random Boolean formulas

We study noisy computation in randomly generated k-ary Boolean formulas. We establish bounds on the noise level above which the results of computation by random formulas are not reliable. This bound is saturated by formulas constructed from a single majority-like gate. We show that these gates can be used to compute any Boolean function reliably below the noise bound.

Analysis of Local Search Methods for 3D Data

In questa tesi sono stati analizzati alcuni metodi di ricerca per dati 3D. Viene illustrata una panoramica generale sul campo della Computer Vision, sullo stato dell’arte dei sensori per l’acquisizione e su alcuni dei formati utilizzati per la descrizione di dati 3D. In seguito è stato fatto un approfondimento sulla 3D Object Recognition dove, oltre ad essere descritto l’intero processo di matching tra Local Features, è stata fatta una focalizzazione sulla fase di detection dei punti salienti. In particolare è stato analizzato un Learned Keypoint detector, basato su tecniche di apprendimento di machine learning. Quest ultimo viene illustrato con l’implementazione di due algoritmi di ricerca di vicini: uno esauriente (K-d tree) e uno approssimato (Radial Search). Sono state riportate infine alcune valutazioni sperimentali in termini di efficienza e velocità del detector implementato con diversi metodi di ricerca, mostrando l’effettivo miglioramento di performance senza una considerabile perdita di accuratezza con la ricerca approssimata.

K-tree : large scale document clustering

We introduce K-tree in an information retrieval context. It is an efficient approximation of the k-means clustering algorithm. Unlike k-means it forms a hierarchy of clusters. It has been extended to address issues with sparse representations. We compare performance and quality to CLUTO using document collections. The K-tree has a low time complexity that is suitable for large document collections. This tree structure allows for efficient disk based implementations where space requirements exceed that of main memory.

Document clustering with K-tree

This paper describes the approach taken to the XML Mining track at INEX 2008 by a group at the Queensland University of Technology. We introduce the K-tree clustering algorithm in an Information Retrieval context by adapting it for document clustering. Many large scale problems exist in document clustering. K-tree scales well with large inputs due to its low complexity. It offers promising results both in terms of efficiency and quality. Document classification was completed using Support Vector Machines.

Random Indexing K-tree

Random Indexing K-tree is the combination of two algorithms suited for large scale document clustering.