Outlier detection with partial information:Application to emergency mapping

This paper, addresses the problem of novelty detection in the case that the observed data is a mixture of a known 'background' process contaminated with an unknown other process, which generates the outliers, or novel observations. The framework we describe here is quite general, employing univariate classification with incomplete information, based on knowledge of the distribution (the 'probability density function', 'pdf') of the data generated by the 'background' process. The relative proportion of this 'background' component (the 'prior' 'background' 'probability), the 'pdf' and the 'prior' probabilities of all other components are all assumed unknown. The main contribution is a new classification scheme that identifies the maximum proportion of observed data following the known 'background' distribution. The method exploits the Kolmogorov-Smirnov test to estimate the proportions, and afterwards data are Bayes optimally separated. Results, demonstrated with synthetic data, show that this approach can produce more reliable results than a standard novelty detection scheme. The classification algorithm is then applied to the problem of identifying outliers in the SIC2004 data set, in order to detect the radioactive release simulated in the 'oker' data set. We propose this method as a reliable means of novelty detection in the emergency situation which can also be used to identify outliers prior to the application of a more general automatic mapping algorithm. © Springer-Verlag 2007.

A combined approach for hiding partial information in RSA

Partial information leakage in deterministic public-key cryptosystems refers to a problem that arises when information about either the plaintext or the key is leaked in subtle ways. Quite a common case is where there are a small number of possible messages that may be sent. An attacker may be able to crack the scheme simply by enumerating all the possible ciphertexts. Two methods are proposed for facing the partial information leakage problem in RSA that incorporate a random element into the encrypted message to increase the number of possible ciphertexts. The resulting scheme is, effectively, an RSA-like cryptosystem which exhibits probabilistic encryption. The first method involves encrypting several similar messages with RSA and then using the Quadratic Residuosity Problem (QRP) to mark the intended one. In this way, an adversary who has correctly guessed two or more of the ciphertexts is still in doubt about which message is the intended one. The cryptographic strength of the combined system is equal to the computational difficulty of factorising a large integer; ideally, this should be feasible. The second scheme uses error-correcting codes for accommodating the random component. The plaintext is processed with an error-correcting code and deliberately corrupted before encryption. The introduced corruption lies within the error-correcting ability of the code, so as to enable the recovery of the original message. The random corruption offers a vast number of possible ciphertexts corresponding to a given plaintext; hence an attacker cannot deduce any useful information from it. The proposed systems are compared to other cryptosystems sharing similar characteristics, in terms of execution time and ciphertext size, so as to determine their practical utility. Finally, parameters which determine the characteristics of the proposed schemes are also examined.

The macroeconomics of price adjustments under information frictions and menu costs

Esta tese se dedica ao estudo de modelos de fixação de preços e suas implicações macroeconômicas. Nos primeiros dois capítulos analiso modelos em que as decisões das firmas sobre seus preços praticados levam em conta custos de menu e de informação. No Capítulo 1 eu estimo tais modelos empregando estatísticas de variações de preços dos Estados Unidos, e concluo que: os custos de informação são significativamente maiores que os custos de menu; os dados claramente favorecem o modelo em que informações sobre condições agregadas são custosas enquanto que as idiossincráticas têm custo zero. No Capítulo 2 investigo as consequências de choques monetários e anúncios de desinflação usando os modelos previamente estimados. Mostro que o grau de não-neutralidade monetária é maior no modelo em que parte da informação é grátis. O Capítulo 3 é um artigo em conjunto com Carlos Carvalho (PUC-Rio) e Antonella Tutino (Federal Reserve Bank of Dallas). No artigo examinamos um modelo de fixação de preços em que firmas estão sujeitas a uma restrição de fluxo de informação do tipo Shannon. Calibramos o modelo e estudamos funções impulso-resposta a choques idiossincráticos e agregados. Mostramos que as firmas vão preferir processar informações agregadas e idiossincráticas conjuntamente ao invés de investigá-las separadamente. Este tipo de processamento gera ajustes de preços mais frequentes, diminuindo a persistência de efeitos reais causados por choques monetários.

Verifying abstract information flow properties in fault tolerant security devices

The verification of information flow properties of security devices is difficult because it involves the analysis of schematic diagrams, artwork, embedded software, etc. In addition, a typical security device has many modes, partial information flow, and needs to be fault tolerant. We propose a new approach to the verification of such devices based upon checking abstract information flow properties expressed as graphs. This approach has been implemented in software, and successfully used to find possible paths of information flow through security devices.

CEO compensation, option incentives and information disclosure

This paper examines the impact of information disclosure on the valuation of CEO options and the incentives created by those options. Prior executive compensation research in the US has made assumptions about key input variables that can affect the calculation of option values and financial incentives. Accordingly, biases may have ensued due to incomplete information disclosure about noncurrent option grants. Using new data on a sample of UK CEOs, we value executive option holdings and incentives for the first time and estimate the levels of distortion created by the less than complete US-style disclosure requirements. We also investigate the levels of distortion in the UK for the minority of companies that choose to reveal only partial information. Our results suggest that there have to date been few economic biases arising from less than complete information disclosure. Furthermore, we demonstrate that researchers using US data, who made reasonable assumptions about the inputs of noncurrent option grants, are unlikely to have made significant errors when calculating CEO financial incentives or option wealth. However, the recent downturn in the US stock market could result in the same assumptions, producing exaggerated incentive estimates in the future.

Tailoring teleportation to the quantum alphabet

We introduce a refinement of the standard continuous variable teleportation measurement and displacement strategies. This refinement makes use of prior knowledge about the target state and the partial information carried by the classical channel when entanglement is nonmaximal. This gives an improvement in the output quality of the protocol. The strategies we introduce could be used in current continuous variable teleportation experiments.

Localitzador gràfic d'adreces IP

Tracta la realització de l'estudi sobre el sistema d'adreçament IP a Internet, a fi d'obtenir una aplicació que ens mostri gràficament sobre un mapa mundial el recorregut des de l'origen fins la destinació. També subministra informació parcial obtinguda al procés de consultes.

Um algoritmo de criptografia de chave pública semanticamente seguro baseado em curvas elípticas

Esta dissertação apresenta o desenvolvimento de um novo algoritmo de criptografia de chave pública. Este algoritmo apresenta duas características que o tornam único, e que foram tomadas como guia para a sua concepção. A primeira característica é que ele é semanticamente seguro. Isto significa que nenhum adversário limitado polinomialmente consegue obter qualquer informação parcial sobre o conteúdo que foi cifrado, nem mesmo decidir se duas cifrações distintas correspondem ou não a um mesmo conteúdo. A segunda característica é que ele depende, para qualquer tamanho de texto claro, de uma única premissa de segurança: que o logaritmo no grupo formado pelos pontos de uma curva elíptica de ordem prima seja computacionalmente intratável. Isto é obtido garantindo-se que todas as diferentes partes do algoritmo sejam redutíveis a este problema. É apresentada também uma forma simples de estendê-lo a fim de que ele apresente segurança contra atacantes ativos, em especial, contra ataques de texto cifrado adaptativos. Para tanto, e a fim de manter a premissa de que a segurança do algoritmo seja unicamente dependente do logaritmo elíptico, é apresentada uma nova função de resumo criptográfico (hash) cuja segurança é baseada no mesmo problema.

Aspectos éticos e legais da competência informacional e a formação do bibliotecário: um estudo com graduandos de biblioteconomia da região sudeste

This study aimed to verify the partial information literacy of librarianship undergraduates with regard to ethical and legal issues related to the use of information and compare with the standards of ACRL (2000). To this end, we applied an electronic questionnaire with 230 students from fifteen schools of Library science in southeastern Brazil. From the results it was realized that the standardization of reference, participants showed a performance below expectations, possibly compromising the recovery of a document by others. In the other questions that addressed issues regarding copyright and intellectual property, the majority of participants have shown that knowledge, behaviors and attitudes appropriate in relation to this aspect.

Automatic (data-based) learning of expressive indicators for machine failure detection : a proposal

Máster Universitario en Sistemas Inteligentes y Aplicaciones Numéricas en Ingeniería (SIANI)

Contribuciones al análisis de problemas supercomplejos de toma de decisiones

Los fundamentos de la Teoría de la Decisión Bayesiana proporcionan un marco coherente en el que se pueden resolver los problemas de toma de decisiones. La creciente disponibilidad de ordenadores potentes está llevando a tratar problemas cada vez más complejos con numerosas fuentes de incertidumbre multidimensionales; varios objetivos conflictivos; preferencias, metas y creencias cambiantes en el tiempo y distintos grupos afectados por las decisiones. Estos factores, a su vez, exigen mejores herramientas de representación de problemas; imponen fuertes restricciones cognitivas sobre los decisores y conllevan difíciles problemas computacionales. Esta tesis tratará estos tres aspectos. En el Capítulo 1, proporcionamos una revisión crítica de los principales métodos gráficos de representación y resolución de problemas, concluyendo con algunas recomendaciones fundamentales y generalizaciones. Nuestro segundo comentario nos lleva a estudiar tales métodos cuando sólo disponemos de información parcial sobre las preferencias y creencias del decisor. En el Capítulo 2, estudiamos este problema cuando empleamos diagramas de influencia (DI). Damos un algoritmo para calcular las soluciones no dominadas en un DI y analizamos varios conceptos de solución ad hoc. El último aspecto se estudia en los Capítulos 3 y 4. Motivado por una aplicación de gestión de embalses, introducimos un método heurístico para resolver problemas de decisión secuenciales. Como muestra resultados muy buenos, extendemos la idea a problemas secuenciales generales y cuantificamos su bondad. Exploramos después en varias direcciones la aplicación de métodos de simulación al Análisis de Decisiones. Introducimos primero métodos de Monte Cario para aproximar el conjunto no dominado en problemas continuos. Después, proporcionamos un método de Monte Cario basado en cadenas de Markov para problemas con información completa con estructura general: las decisiones y las variables aleatorias pueden ser continuas, y la función de utilidad puede ser arbitraria. Nuestro esquema es aplicable a muchos problemas modelizados como DI. Finalizamos con un capítulo de conclusiones y problemas abiertos.---ABSTRACT---The foundations of Bayesian Decisión Theory provide a coherent framework in which decisión making problems may be solved. With the advent of powerful computers and given the many challenging problems we face, we are gradually attempting to solve more and more complex decisión making problems with high and multidimensional uncertainty, múltiple objectives, influence of time over decisión tasks and influence over many groups. These complexity factors demand better representation tools for decisión making problems; place strong cognitive demands on the decison maker judgements; and lead to involved computational problems. This thesis will deal with these three topics. In recent years, many representation tools have been developed for decisión making problems. In Chapter 1, we provide a critical review of most of them and conclude with recommendations and generalisations. Given our second query, we could wonder how may we deal with those representation tools when there is only partial information. In Chapter 2, we find out how to deal with such a problem when it is structured as an influence diagram (ID). We give an algorithm to compute nondominated solutions in ID's and analyse several ad hoc solution concepts.- The last issue is studied in Chapters 3 and 4. In a reservoir management case study, we have introduced a heuristic method for solving sequential decisión making problems. Since it shows very good performance, we extend the idea to general problems and quantify its goodness. We explore then in several directions the application of simulation based methods to Decisión Analysis. We first introduce Monte Cario methods to approximate the nondominated set in continuous problems. Then, we provide a Monte Cario Markov Chain method for problems under total information with general structure: decisions and random variables may be continuous, and the utility function may be arbitrary. Our scheme is applicable to many problems modeled as IDs. We conclude with discussions and several open problems.

Continuous-variable quantum-state sharing via quantum disentanglement

Quantum-state sharing is a protocol where perfect reconstruction of quantum states is achieved with incomplete or partial information in a multipartite quantum network. Quantum-state sharing allows for secure communication in a quantum network where partial information is lost or acquired by malicious parties. This protocol utilizes entanglement for the secret-state distribution and a class of quantum disentangling protocols for the state reconstruction. We demonstrate a quantum-state sharing protocol in which a tripartite entangled state is used to encode and distribute a secret state to three players. Any two of these players can collaborate to reconstruct the secret state, while individual players obtain no information. We investigate a number of quantum disentangling processes and experimentally demonstrate quantum-state reconstruction using two of these protocols. We experimentally measure a fidelity, averaged over all reconstruction permutations, of F=0.73 +/- 0.02. A result achievable only by using quantum resources.

On optimal completions of incomplete pairwise comparison matrices

An important variant of a key problem for multi-attribute decision making is considered. We study the extension of the pairwise comparison matrix to the case when only partial information is available: for some pairs no comparison is given. It is natural to define the inconsistency of a partially filled matrix as the inconsistency of its best, completely filled completion. We study here the uniqueness problem of the best completion for two weighting methods, the Eigen-vector Method and the Logarithmic Least Squares Method. In both settings we obtain the same simple graph theoretic characterization of the uniqueness. The optimal completion will be unique if and only if the graph associated with the partially defined matrix is connected. Some numerical experiences are discussed at the end of the paper.

Incomplete pairwise comparison matrices and weighting methods

A special class of preferences, given by a directed acyclic graph, is considered. They are represented by incomplete pairwise comparison matrices as only partial information is available: for some pairs no comparison is given in the graph. A weighting method satisfies the property linear order preservation if it always results in a ranking such that an alternative directly preferred to another does not have a lower rank. We study whether two procedures, the Eigenvector Method and the Logarithmic Least Squares Method meet this axiom. Both weighting methods break linear order preservation, moreover, the ranking according to the Eigenvector Method depends on the incomplete pairwise comparison representation chosen.

Adaptive Crowd Algorithms for Open-Ended Problems

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