Discriminative training of the hidden vector state model for semantic parsing

In this paper, we discuss how discriminative training can be applied to the hidden vector state (HVS) model in different task domains. The HVS model is a discrete hidden Markov model (HMM) in which each HMM state represents the state of a push-down automaton with a finite stack size. In previous applications, maximum-likelihood estimation (MLE) is used to derive the parameters of the HVS model. However, MLE makes a number of assumptions and unfortunately some of these assumptions do not hold. Discriminative training, without making such assumptions, can improve the performance of the HVS model by discriminating the correct hypothesis from the competing hypotheses. Experiments have been conducted in two domains: the travel domain for the semantic parsing task using the DARPA Communicator data and the Air Travel Information Services (ATIS) data and the bioinformatics domain for the information extraction task using the GENIA corpus. The results demonstrate modest improvements of the performance of the HVS model using discriminative training. In the travel domain, discriminative training of the HVS model gives a relative error reduction rate of 31 percent in F-measure when compared with MLE on the DARPA Communicator data and 9 percent on the ATIS data. In the bioinformatics domain, a relative error reduction rate of 4 percent in F-measure is achieved on the GENIA corpus.

Relaxing claims:coping with uncertainty while evaluating assumptions at run time

Self-adaptation enables software systems to respond to changing environmental contexts that may not be fully understood at design time. Designing a dynamically adaptive system (DAS) to cope with this uncertainty is challenging, as it is impractical during requirements analysis and design time to anticipate every environmental condition that the DAS may encounter. Previously, the RELAX language was proposed to make requirements more tolerant to environmental uncertainty, and Claims were applied as markers of uncertainty that document how design assumptions affect goals. This paper integrates these two techniques in order to assess the validity of Claims at run time while tolerating minor and unanticipated environmental conditions that can trigger adaptations. We apply the proposed approach to the dynamic reconfiguration of a remote data mirroring network that must diffuse data while minimizing costs and exposure to data loss. Results show RELAXing Claims enables a DAS to reduce adaptation costs. © 2012 Springer-Verlag.

Study of Queueing Behaviour in IP Buffers

It is unquestioned that the importance of IP network will further increase and that it will serve as a platform for more and more services, requiring different types and degrees of service quality. Modern architectures and protocols are being standardized, which aims at guaranteeing the quality of service delivered to users. In this paper, we investigate the queueing behaviour found in IP output buffers. This queueing increases because multiple streams of packets with different length are being multiplexed together. We develop balance equations for the state of the system, from which we derive packet loss and delay results. To analyze these types of behaviour, we study the discrete-time version of the “classical” queue model M/M/1/k called Geo/Gx/1/k, where Gx denotes a different packet length distribution defined on a range between a minimum and maximum value.

Modelling stock volatilities during financial crises:a time varying coefficient approach

We examine how the most prevalent stochastic properties of key financial time series have been affected during the recent financial crises. In particular we focus on changes associated with the remarkable economic events of the last two decades in the volatility dynamics, including the underlying volatility persistence and volatility spillover structure. Using daily data from several key stock market indices, the results of our bivariate GARCH models show the existence of time varying correlations as well as time varying shock and volatility spillovers between the returns of FTSE and DAX, and those of NIKKEI and Hang Seng, which became more prominent during the recent financial crisis. Our theoretical considerations on the time varying model which provides the platform upon which we integrate our multifaceted empirical approaches are also of independent interest. In particular, we provide the general solution for time varying asymmetric GARCH specifications, which is a long standing research topic. This enables us to characterize these models by deriving, first, their multistep ahead predictors, second, the first two time varying unconditional moments, and third, their covariance structure.

A study of stock market linkages between the US and frontier countries

My dissertation investigates the financial linkages and transmission of economic shocks between the US and the smallest emerging markets (frontier markets). The first chapter sets up an empirical model that examines the impact of US market returns and conditional volatility on the returns and conditional volatilities of twenty-one frontier markets. The model is estimated via maximum likelihood; utilizes the GARCH model of errors, and is applied to daily country data from the MSCI Barra. We find limited, but statistically significant exposure of Frontier markets to shocks from the US. Our results suggest that it is not the lagged US market returns that have impact; rather it is the expected US market returns that influence frontier market returns The second chapter sets up an empirical time-varying parameter (TVP) model to explore the time-variation in the impact of mean US returns on mean Frontier market returns. The model utilizes the Kalman filter algorithm as well as the GARCH model of errors and is applied to daily country data from the MSCI Barra. The TVP model detects statistically significant time-variation in the impact of US returns and low, but statistically and quantitatively important impact of US market conditional volatility. The third chapter studies the risk-return relationship in twenty Frontier country stock markets by setting up an international version of the intertemporal capital asset pricing model. The systematic risk in this model comes from covariance of Frontier market stock index returns with world returns. Both the systematic risk and risk premium are time-varying in our model. We also incorporate own country variances as additional determinants of Frontier country returns. Our results suggest statistically significant impact of both world and own country risk in explaining Frontier country returns. Time-variation in the world risk premium is also found to be statistically significant for most Frontier market returns. However, own country risk is found to be quantitatively more important.

Evidence of the heterogeneous market hypothesis using wavelet multi -resolution analysis

In finance literature many economic theories and models have been proposed to explain and estimate the relationship between risk and return. Assuming risk averseness and rational behavior on part of the investor, the models are developed which are supposed to help in forming efficient portfolios that either maximize (minimize) the expected rate of return (risk) for a given level of risk (rates of return). One of the most used models to form these efficient portfolios is the Sharpe's Capital Asset Pricing Model (CAPM). In the development of this model it is assumed that the investors have homogeneous expectations about the future probability distribution of the rates of return. That is, every investor assumes the same values of the parameters of the probability distribution. Likewise financial volatility homogeneity is commonly assumed, where volatility is taken as investment risk which is usually measured by the variance of the rates of return. Typically the square root of the variance is used to define financial volatility, furthermore it is also often assumed that the data generating process is made of independent and identically distributed random variables. This again implies that financial volatility is measured from homogeneous time series with stationary parameters. In this dissertation, we investigate the assumptions of homogeneity of market agents and provide evidence for the case of heterogeneity in market participants' information, objectives, and expectations about the parameters of the probability distribution of prices as given by the differences in the empirical distributions corresponding to different time scales, which in this study are associated with different classes of investors, as well as demonstrate that statistical properties of the underlying data generating processes including the volatility in the rates of return are quite heterogeneous. In other words, we provide empirical evidence against the traditional views about homogeneity using non-parametric wavelet analysis on trading data, The results show heterogeneity of financial volatility at different time scales, and time-scale is one of the most important aspects in which trading behavior differs. In fact we conclude that heterogeneity as posited by the Heterogeneous Markets Hypothesis is the norm and not the exception.

Um controlador dual subótimo com fator de peso variável no tempo

This work concerns a refinement of a suboptimal dual controller for discrete time systems with stochastic parameters. The dual property means that the control signal is chosen so that estimation of the model parameters and regulation of the output signals are optimally balanced. The control signal is computed in such a way so as to minimize the variance of output around a reference value one step further, with the addition of terms in the loss function. The idea is add simple terms depending on the covariance matrix of the parameter estimates two steps ahead. An algorithm is used for the adaptive adjustment of the adjustable parameter lambda, for each step of the way. The actual performance of the proposed controller is evaluated through a Monte Carlo simulations method.

Distributed Extremum-Seeking Control with Applications to High-Altitude Balloons

The real-time optimization of large-scale systems is a difficult problem due to the need for complex models involving uncertain parameters and the high computational cost of solving such problems by a decentralized approach. Extremum-seeking control (ESC) is a model-free real-time optimization technique which can estimate unknown parameters and can optimize nonlinear time-varying systems using only a measurement of the cost function to be minimized. In this thesis, we develop a distributed version of extremum-seeking control which allows large-scale systems to be optimized without models and with minimal computing power. First, we develop a continuous-time distributed extremum-seeking controller. It has three main components: consensus, parameter estimation, and optimization. The consensus provides each local controller with an estimate of the cost to be minimized, allowing them to coordinate their actions. Using this cost estimate, parameters for a local input-output model are estimated, and the cost is minimized by following a gradient descent based on the estimate of the gradient. Next, a similar distributed extremum-seeking controller is developed in discrete-time. Finally, we consider an interesting application of distributed ESC: formation control of high-altitude balloons for high-speed wireless internet. These balloons must be steered into a favourable formation where they are spread out over the Earth and provide coverage to the entire planet. Distributed ESC is applied to this problem, and is shown to be effective for a system of 1200 ballons subjected to realistic wind currents. The approach does not require a wind model and uses a cost function based on a Voronoi partition of the sphere. Distributed ESC is able to steer balloons from a few initial launch sites into a formation which provides coverage to the entire Earth and can maintain a similar formation as the balloons move with the wind around the Earth.

A Bias Compensated Cross-Relation approach to Thermocouple Characterisation

The measurement of fast changing temperature fluctuations is a challenging problem due to the inherent limited bandwidth of temperature sensors. This results in a measured signal that is a lagged and attenuated version of the input. Compensation can be performed provided an accurate, parameterised sensor model is available. However, to account for the in influence of the measurement environment and changing conditions such as gas velocity, the model must be estimated in-situ. The cross-relation method of blind deconvolution is one approach for in-situ characterisation of sensors. However, a drawback with the method is that it becomes positively biased and unstable at high noise levels. In this paper, the cross-relation method is cast in the discrete-time domain and a bias compensation approach is developed. It is shown that the proposed compensation scheme is robust and yields unbiased estimates with lower estimation variance than the uncompensated version. All results are verified using Monte-Carlo simulations.

A Bias Compensated Cross-Relation approach to Thermocouple Characterisation

The measurement of fast changing temperature fluctuations is a challenging problem due to the inherent limited bandwidth of temperature sensors. This results in a measured signal that is a lagged and attenuated version of the input. Compensation can be performed provided an accurate, parameterised sensor model is available. However, to account for the influence of the measurement environment and changing conditions such as gas velocity, the model must be estimated in-situ. The cross-relation method of blind deconvolution is one approach for in-situ characterisation of sensors. However, a drawback with the method is that it becomes positively biased and unstable at high noise levels. In this paper, the cross-relation method is cast in the discrete-time domain and a bias compensation approach is developed. It is shown that the proposed compensation scheme is robust and yields unbiased estimates with lower estimation variance than the uncompensated version. All results are verified using Monte-Carlo simulations.

Managing Rentals with Usage-Based Loss

Motivated by new and innovative rental business models, this paper develops a novel discrete-time model of a rental operation with random loss of inventory due to customer use. The inventory level is chosen before the start of a finite rental season, and customers not immediately served are lost. Our analysis framework uses stochastic comparisons of sample paths to derive structural results that hold under good generality for demands, rental durations, and rental unit lifetimes. Considering different \recirculation" rules | i.e., which rental unit to choose to meet each demand | we prove the concavity of the expected profit function and identify the optimal recirculation rule. A numerical study clarifies when considering rental unit loss and recirculation rules matters most for the inventory decision: Accounting for rental unit loss can increase the expected profit by 7% for a single season and becomes even more important as the time horizon lengthens. We also observe that the optimal inventory level in response to increasing loss probability is non-monotonic. Finally, we show that choosing the optimal recirculation rule over another simple policy allows more rental units to be profitably added, and the profit-maximizing service level increases by up to 6 percentage points.

Evolutionary dynamics under various modes of reproduction

Thesis (Ph.D.)--University of Washington, 2016-08

Age-Specific Fertility Dynamics: Sub-Saharan African Fertility in a Global Context

Thesis (Ph.D.)--University of Washington, 2016-08

Structural Time Series Modeling: an Application to Environmental Variables

A structural time series model is one which is set up in terms of components which have a direct interpretation. In this paper, the discussion focuses on the dynamic modeling procedure based on the state space approach (associated to the Kalman filter), in the context of surface water quality monitoring, in order to analyze and evaluate the temporal evolution of the environmental variables, and thus identify trends or possible changes in water quality (change point detection). The approach is applied to environmental time series: time series of surface water quality variables in a river basin. The statistical modeling procedure is applied to monthly values of physico- chemical variables measured in a network of 8 water monitoring sites over a 15-year period (1999-2014) in the River Ave hydrological basin located in the Northwest region of Portugal.

Correction for guessing in choice reaction time

Additional theoretical and experimental results are presented for a choice reaction time performance model described by Oilman (1966). A formula is given for estimating the latency distribution of true recognition responses from the results of a single session; the estimate is invariant with respect to changes in the proportion of “guess” responses and with respect to fluctuations in the latency distribution of guesses. © 1967, Psychonomic Press. All rights reserved.