Stanford Goldman. / "December 15, 1947." "Presented at the National Electronics Conference, Chicago, Illinois, November, 1947." / Includes bibliographical references. / Army Signal Corps Contract No. W-36-039 sc-32037.

This memo describes the initial results of a project to create a self-supervised algorithm for learning object segmentation from video data. Developmental psychology and computational experience have demonstrated that the motion segmentation of objects is a simpler, more primitive process than the detection of object boundaries by static image cues. Therefore, motion information provides a plausible supervision signal for learning the static boundary detection task and for evaluating performance on a test set. A video camera and previously developed background subtraction algorithms can automatically produce a large database of motion-segmented images for minimal cost. The purpose of this work is to use the information in such a database to learn how to detect the object boundaries in novel images using static information, such as color, texture, and shape. This work was funded in part by the Office of Naval Research contract #N00014-00-1-0298, in part by the Singapore-MIT Alliance agreement of 11/6/98, and in part by a National Science Foundation Graduate Student Fellowship.

We seek to both detect and segment objects in images. To exploit both local image data as well as contextual information, we introduce Boosted Random Fields (BRFs), which uses Boosting to learn the graph structure and local evidence of a conditional random field (CRF). The graph structure is learned by assembling graph fragments in an additive model. The connections between individual pixels are not very informative, but by using dense graphs, we can pool information from large regions of the image; dense models also support efficient inference. We show how contextual information from other objects can improve detection performance, both in terms of accuracy and speed, by using a computational cascade. We apply our system to detect stuff and things in office and street scenes.

Exponential Random Graph Models (ERGMs) have been developed for fitting social network data on both static and dynamic levels. However, the lack of large sample asymptotic properties makes it inadequate in assessing the goodness-of-fit of these ERGMs. Simulation-based goodness-of-fit plots were proposed by Hunter et al (2006), comparing the structured statistics of observed network with those of corresponding simulated networks. In this research, we propose an improved approach to assess the goodness of fit of ERGMs. Our method is shown to improve the existing graphical techniques. We also propose a simulation based test statistic with which the model comparison can be easily achieved. / Biostatistics

No description available.

No description available.

No description available.

I uppsatsen undersöks effektiviteten av de tre nordiska aktieindexterminerna OMXS30, OBX och OMXC20 vars underliggande index representerar den svenska, norska respektive den danska aktiemarknaden. Analysen baseras på den svaga formen av den effektiva marknadshypotesen och den närbesläktade random walk hypotesen. Aktieindexterminerna undersöks under perioden januari 1997 till december 2008 samt under perioder då den nordiska marknaden karaktäriseras av bull och bear perioder. Testresultaten av Augmented Dickey-Fuller (ADF) samt Kwiatkowski, Phillips, Schmidt och Shin (KPSS) testet tyder på att aktieindexterminerna följer en random walk och att nordiska aktieindexterminer är effektiva under den undersökta perioden. Då testen utförs för de kortare bull och bear perioderna erhålls motsägelsefulla resultat vilket medför att slutsatser om huruvida aktieindexterminerna är effektiva under dessa perioder ej kan dras.

Denna uppsats undersöker huruvida ett antal aktiemarknader på Balkan tillfredställer den svaga formen av marknadseffektivitet under perioden 2003-2007. Svag marknadseffektivitet innebär att priset på en finansiell tillgång reflekterar all historisk prisinformation. Om en marknad är svagt effektiv leder analys av historisk kursinformation inte till överavkastning eftersom all historisk information redan är diskonterad av marknaden. För att pröva den svaga formen av effektivitet för Kroatien, Makedonien, Montenegro, Serbien och Slovenien genomförs olika statistiska tester såsom autokorrelationstest, runs test, varianskvottest och regressionsanalys. Samtliga tester i studien genomförs på dagsavkastningar. Resultatet av undersökningen visar att samtliga avkastningsserier lider av positiv autokorrelation, vilket innebär att den historiska avkastningen i viss mån kan användas för att predicera den framtida avkastningen. Detta behöver dock inte betyda avvikelse från den effektiva marknadsmodellen då den ekonomiska signifikansen är relativt låg med en genomsnittlig förklarningsgrad under fem procent.

Nitriding and oxidising treatments of alloys for toolingpurposes have been investigated both experimentally andtheoretically. Specimens prepared by a combination of differentprocessing steps are investigated by light-optical, scanningand transmission electron microscopy and hardness measurments.The bounds of carbon and nitrogen concentration profiles duringnitriding are obtained from simulations using the DICTRA code.As a result, a new processing route to produce oxide-dispersediron-base alloys with large volume fraction of carbides issuggested. On the theoretical side the methods based on the random walktechnique have been investigated and developed further. The newmethods have been applied to study a vide range of phenomenae.g. internal oxidation, diffusion in inhomogeneous media,phase transformations and formation of porosity due to theKirkendall effect in welded joints. Probability distributionsare introduced to replace random number generators in order toincrease computational efficiency. A general method to simulatediffusional phase transformations in multicomponent systems isdeveloped and applied to ternary alloys. / QC 20100506