Combining Information from Multiple Sources in Bayesian Modeling

Surveys can collect important data that inform policy decisions and drive social science research. Large government surveys collect information from the U.S. population on a wide range of topics, including demographics, education, employment, and lifestyle. Analysis of survey data presents unique challenges. In particular, one needs to account for missing data, for complex sampling designs, and for measurement error. Conceptually, a survey organization could spend lots of resources getting high-quality responses from a simple random sample, resulting in survey data that are easy to analyze. However, this scenario often is not realistic. To address these practical issues, survey organizations can leverage the information available from other sources of data. For example, in longitudinal studies that suffer from attrition, they can use the information from refreshment samples to correct for potential attrition bias. They can use information from known marginal distributions or survey design to improve inferences. They can use information from gold standard sources to correct for measurement error.

This thesis presents novel approaches to combining information from multiple sources that address the three problems described above.

The first method addresses nonignorable unit nonresponse and attrition in a panel survey with a refreshment sample. Panel surveys typically suffer from attrition, which can lead to biased inference when basing analysis only on cases that complete all waves of the panel. Unfortunately, the panel data alone cannot inform the extent of the bias due to attrition, so analysts must make strong and untestable assumptions about the missing data mechanism. Many panel studies also include refreshment samples, which are data collected from a random sample of new

individuals during some later wave of the panel. Refreshment samples offer information that can be utilized to correct for biases induced by nonignorable attrition while reducing reliance on strong assumptions about the attrition process. To date, these bias correction methods have not dealt with two key practical issues in panel studies: unit nonresponse in the initial wave of the panel and in the

refreshment sample itself. As we illustrate, nonignorable unit nonresponse

can significantly compromise the analyst's ability to use the refreshment samples for attrition bias correction. Thus, it is crucial for analysts to assess how sensitive their inferences---corrected for panel attrition---are to different assumptions about the nature of the unit nonresponse. We present an approach that facilitates such sensitivity analyses, both for suspected nonignorable unit nonresponse

in the initial wave and in the refreshment sample. We illustrate the approach using simulation studies and an analysis of data from the 2007-2008 Associated Press/Yahoo News election panel study.

The second method incorporates informative prior beliefs about

marginal probabilities into Bayesian latent class models for categorical data.

The basic idea is to append synthetic observations to the original data such that

(i) the empirical distributions of the desired margins match those of the prior beliefs, and (ii) the values of the remaining variables are left missing. The degree of prior uncertainty is controlled by the number of augmented records. Posterior inferences can be obtained via typical MCMC algorithms for latent class models, tailored to deal efficiently with the missing values in the concatenated data.

We illustrate the approach using a variety of simulations based on data from the American Community Survey, including an example of how augmented records can be used to fit latent class models to data from stratified samples.

The third method leverages the information from a gold standard survey to model reporting error. Survey data are subject to reporting error when respondents misunderstand the question or accidentally select the wrong response. Sometimes survey respondents knowingly select the wrong response, for example, by reporting a higher level of education than they actually have attained. We present an approach that allows an analyst to model reporting error by incorporating information from a gold standard survey. The analyst can specify various reporting error models and assess how sensitive their conclusions are to different assumptions about the reporting error process. We illustrate the approach using simulations based on data from the 1993 National Survey of College Graduates. We use the method to impute error-corrected educational attainments in the 2010 American Community Survey using the 2010 National Survey of College Graduates as the gold standard survey.

Incomplete DJH rearrangements as a novel tumor target for minimal residual disease quantitation in multiple myeloma using real-time PCR.

The hypervariable regions of immunoglobulin heavy-chain (IgH) rearrangements provide a specific tumor marker in multiple myeloma (MM). Recently, real-time PCR assays have been developed in order to quantify the number of tumor cells after treatment. However, these strategies are hampered by the presence of somatic hypermutation (SH) in VDJH rearrangements from multiple myeloma (MM) patients, which causes mismatches between primers and/or probes and the target, leading to a nonaccurate quantification of tumor cells. Our group has recently described a 60% incidence of incomplete DJH rearrangements in MM patients, with no or very low rates of SH. In this study, we compare the efficiency of a real-time PCR approach for the analysis of both complete and incomplete IgH rearrangements in eight MM patients using only three JH consensus probes. We were able to design an allele-specific oligonucleotide for both the complete and incomplete rearrangement in all patients. DJH rearrangements fulfilled the criteria of effectiveness for real-time PCR in all samples (ie no unspecific amplification, detection of less than 10 tumor cells within 10(5) polyclonal background and correlation coefficients of standard curves higher than 0.98). By contrast, only three out of eight VDJH rearrangements fulfilled these criteria. Further analyses showed that the remaining five VDJH rearrangements carried three or more somatic mutations in the probe and primer sites, leading to a dramatic decrease in the melting temperature. These results support the use of incomplete DJH rearrangements instead of complete somatically mutated VDJH rearrangements for investigation of minimal residual disease in multiple myeloma.

Secure 5G Wireless Communications: A Joint Relay Selection and Wireless Power Transfer Approach

In this paper, we investigate the secrecy performance of an energy harvesting relay system, where a legitimate source communicates with a legitimate destination via the assistance of multiple trusted relays. In the considered system, the source and relays deploy the time-switching-based radio frequency energy harvesting technique to harvest energy from a multi-antenna beacon. Different antenna selection and relay selection schemes are applied to enhance the security of the system. Specifically, two relay selection schemes based on the partial and full knowledge of channel state information, i.e., optimal relay selection and partial relay selection, and two antenna selection schemes for harvesting energy at source and relays, i.e., maximizing energy harvesting channel for the source and maximizing energy harvesting channel for the selected relay, are proposed. The exact and asymptotic expressions of secrecy outage probability in these schemes are derived. We demonstrate that applying relay selection approaches in the considered energy harvesting system can enhance the security performance. In particular, optimal relay selection scheme outperforms partial relay selection scheme and achieves full secrecy diversity order, regardless of energy harvesting scenarios.

A Two-Stage Online Approach for Collaborative Multi-Agent Planning under Uncertainty

In a team of multiple agents, the pursuance of a common goal is a defining characteristic. Since agents may have different capabilities, and effects of actions may be uncertain, a common goal can generally only be achieved through a careful cooperation between the different agents. In this work, we propose a novel two-stage planner that combines online planning at both team level and individual level through a subgoal delegation scheme. The proposal brings the advantages of online planning approaches to the multi-agent setting. A number of modifications are made to a classical UCT approximate algorithm to (i) adapt it to the application domains considered, (ii) reduce the branching factor in the underlying search process, and (iii) effectively manage uncertain information of action effects by using information fusion mechanisms. The proposed online multi-agent planner reduces the cost of planning and decreases the temporal cost of reaching a goal, while significantly increasing the chance of success of achieving the common goal. 

Joint Analysis of Multiple Algorithms and Performance Measures

There has been an increasing interest in the development of new methods using Pareto optimality to deal with multi-objective criteria (for example, accuracy and time complexity). Once one has developed an approach to a problem of interest, the problem is then how to compare it with the state of art. In machine learning, algorithms are typically evaluated by comparing their performance on different data sets by means of statistical tests. Standard tests used for this purpose are able to consider jointly neither performance measures nor multiple competitors at once. The aim of this paper is to resolve these issues by developing statistical procedures that are able to account for multiple competing measures at the same time and to compare multiple algorithms altogether. In particular, we develop two tests: a frequentist procedure based on the generalized likelihood-ratio test and a Bayesian procedure based on a multinomial-Dirichlet conjugate model. We further extend them by discovering conditional independences among measures to reduce the number of parameters of such models, as usually the number of studied cases is very reduced in such comparisons. Data from a comparison among general purpose classifiers is used to show a practical application of our tests.

Complexity of risk: Mixed methods approach to understanding youth risk and insecurity in post-conflict settings

In settings of intergroup conflict, identifying contextually-relevant risk factors for youth development in an important task. In Vukovar, Croatia, a city devastated during the war in former Yugoslavia, ethno-political tensions remain. The current study utilized a mixed method approach to identify two salient community-level risk factors (ethnic tension and general antisocial behavior) and related emotional insecurity responses (ethnic and non-ethnic insecurity) among youth in Vukovar. In Study 1, focus group discussions (N=66) with mother, fathers, and adolescents 11 to 15-years-old were analyzed using the Constant Comparative Method, revealing two types of risk and insecurity responses. In Study 2, youth (N=227, 58% male, M=15.88 SD=1.12 years old) responded to quantitative scales developed from the focus groups; discriminate validity was demonstrated and path analyses established predictive validity between each type of risk and insecurity. First, community ethnic tension (i.e., threats related to war/ethnic identity) significantly predicted ethnic insecurity for all youth (β=.41, p<.001). Second, experience with community antisocial behavior (i.e., general crime found in any context) predicted non-ethnic community insecurity for girls (β=.32, p<.05), but not for boys. These findings are the first to show multiple forms of emotional insecurity at the community level; implications for future research are discussed.

Multiple face detection at different resolutions for perceptual user interfaces

[EN]This paper describes in detail a real-time multiple face detection system for video streams. The system adds to the good performance provided by a window shift approach, the combination of different cues available in video streams due to temporal coherence. The results achieved by this combined solution outperform the basic face detector obtaining a 98% success rate for around 27000 images, providing additionally eye detection and a relation between the successive detections in time by means of detection threads.

Determining day length and temperature regulation of flowering: a molecular and modelling approach

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

Random Access over Multi-Packet Reception Channels: A Mean-Field Approach

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

Development of a transitioning approach to reduce surface water volumes in combined sewer systems

The overarching goal of this research is to establish a successful forum for a transition from the existing paradigm of managing wastewater infrastructure to a more sustainable paradigm that achieves a more efficient utilisation of wastewater assets. A transitioning approach to support a more efficient utilisation of surface water and wastewater assets and infrastructure is proposed and developed. The determined transitioning approach possesses key stages namely developing the arena, developing the agenda, case study, and monitoring. The case study stage investigates a drainage utility identifying their improvement drivers, the removal of surface water through detailed drainage modelling and the financial examination of the costs incurred under the various scenarios conducted. Understanding the implications of removing/attenuating surface water from the network is improved through obtaining data by detailed drainage modelling. Infoworks software is used to investigate and assess the current and future operational scenarios of a wastewater system operating over one calendar year. Modelling scenarios were conducted removing surface water from selected areas focusing on the volumes requiring pumping and durations of pumping station(s) operation prior to treatment during storm conditions. The financial implication of removing surface water in combined sewer systems is examined in three main components. Firstly the costs of electricity incurred at the single sewage pumping station (SPS) investigated during the various scenarios modelled require to be addressed. Secondly the costs to retrofit sustainable urban drainage system (SUDS) solutions needs to be identified. Thirdly the implications of removing surface water for the drainage utility at the national level and the potential saving for householder’s committing to a surface water disconnection rebate scheme. When addressed at the macro level i.e., with over 2,100 pumping stations, some operating in sequence and contained within one drainage utility annually treating 315,360 megalitres the significance of the same multiple quantifiable and intangible benefits becomes amplified. The research aims, objectives and findings are presented to the identified and convened stakeholders. The transitioning approach developed encourages positive discourse between stakeholders. The level of success of the transitioning approach determined is then tested using a quantitative methodology through the completion of questionnaires. From the questionnaires completed the respondents unanimously agreed that surface water flows should be removed as well as reduced from the combined sewer system. The respondents agreed that the removal of surface water from a typical combined sewer system is justified by applying a transitioning approach focusing on the energy consumption required to pump increased volumes during storm events. This response is significant based upon the economic evidence and is contrary to the respondents previous position that finance was their most influencing factor. When provided with other potentially available benefits the respondents were even more supportive of the justification to remove surface water from the combined sewer system. The combined findings of the work presented in this thesis provide further justification that the transitioning approach applied to the removal of surface water from a typical combined sewer system, as determined in this research has been successful.

Optimization of a multielectrode array (MEA)-based approach to study the impact of Aβ on the SH-SY5Y cell line

The human brain stores, integrates, and transmits information recurring to millions of neurons, interconnected by countless synapses. Though neurons communicate through chemical signaling, information is coded and conducted in the form of electrical signals. Neuroelectrophysiology focus on the study of this type of signaling. Both intra and extracellular approaches are used in research, but none holds as much potential in high-throughput screening and drug discovery, as extracellular recordings using multielectrode arrays (MEAs). MEAs measure neuronal activity, both in vitro and in vivo. Their key advantage is the capability to record electrical activity at multiple sites simultaneously. Alzheimer’s disease (AD) is the most common neurodegenerative disease and one of the leading causes of death worldwide. It is characterized by neurofibrillar tangles and aggregates of amyloid-β (Aβ) peptides, which lead to the loss of synapses and ultimately neuronal death. Currently, there is no cure and the drugs available can only delay its progression. In vitro MEA assays enable rapid screening of neuroprotective and neuroharming compounds. Therefore, MEA recordings are of great use in both AD basic and clinical research. The main aim of this thesis was to optimize the formation of SH-SY5Y neuronal networks on MEAs. These can be extremely useful for facilities that do not have access to primary neuronal cultures, but can also save resources and facilitate obtaining faster high-throughput results to those that do. Adhesion-mediating compounds proved to impact cell morphology, viability and exhibition of spontaneous electrical activity. Moreover, SH-SY5Y cells were successfully differentiated and demonstrated acute effects on neuronal function after Aβ addition. This effect on electrical signaling was dependent on Aβ oligomers concentration. The results here presented allow us to conclude that the SH-SY5Y cell line can be successfully differentiated in properly coated MEAs and be used for assessing acute Aβ effects on neuronal signaling.

A network approach to overcoming barriers to market engagement for SMEs in energy efficiency initiatives such as the Green Deal

The Green Deal (GD) was launched in 2013 by the UK Government as a market-led scheme to encourage uptake of energy efficiency measures in the UK and create green sector jobs. The scheme closed in July 2015 after 30 months due to government concerns over low uptake and industry standards but additional factors potentially contributed to its failure such as poor scheme design and lack of understanding of the customer and supply chain journey. We explore the role of key delivery agents of GD services, specifically SMEs, and we use the LoCal-Net project as a case study to examine the use of networks to identify and reduce barriers to SME market engagement. We find that SMEs experienced multiple barriers to interaction with the GD such as lack of access to information, training, and confusion over delivery of the scheme but benefited from interaction with the network to access information, improve understanding of the scheme, increasing networking opportunities and forming new business models and partnerships to reduce risk. The importance of SMEs as delivery agents and their role in the design of market-led schemes such as the GD are discussed with recommendations for improving SME engagement in green sector initiatives.

Multiple endocrine neoplasia, the old and the new: a mini review

Multiple endocrine neoplasia syndromes have since been classified as types 1 and 2, each with specific phenotypic patterns. MEN1 is usually associated with pituitary, parathyroid and paraneoplastic neuroendocrine tumours. The hallmark of MEN2 is a very high lifetime risk of developing medullary thyroid carcinoma (MTC) more than 95% in untreated patients. Three clinical subtypesdMEN2A, MEN2B, and familial MTC (FMTC) have been defined based on the risk of pheochromocytoma, hyperparathyroidism, and the presence or absence of characteristic physical features). MEN2 occurs as a result of germline activating missense mutations of the RET (REarranged during Transfection) proto-oncogene. MEN2-associated mutations are almost always located in exons 10, 11, or 13 through 16. Strong genotype-phenotype correlations exist with respect to clinical subtype, age at onset, and aggressiveness of MTC in MEN2. These are used to determine the age at which prophylactic thyroidectomy should occur and whether screening for pheochromocytoma or hyperparathyroidism is necessary. Specific RET mutations can also impact management in patients presenting with apparently sporadic MTC. Therefore, genetic testing should be performed before surgical intervention in all patients diagnosed with MTC. Recently, Pellegata et al. have reported that germline mutations in CDKN1B can predispose to the development of multiple endocrine tumours in both rats and humans and this new MEN syndrome is named MENX and MEN4, respectively. CDKN1B. A recent report showed that in sporadic MTC, CDKN1B V109G polymorphism correlates with a more favorable disease progression than the wild-type allele and might be considered a new promising prognostic marker. New insights on MEN syndrome pathogenesis and related inherited endocrine disorders are of particular interest for an adequate surgical and therapeutic approach.

Rice-Infecting Pseudomonas Genomes Are Highly Accessorized and Harbor Multiple Putative Virulence Mechanisms to Cause Sheath Brown Rot

Sheath rot complex and seed discoloration in rice involve a number of pathogenic bacteria that cannot be associated with distinctive symptoms. These pathogens can easily travel on asymptomatic seeds and therefore represent a threat to rice cropping systems. Among the rice-infecting Pseudomonas, P. fuscovaginae has been associated with sheath brown rot disease in several rice growing areas around the world. The appearance of a similar Pseudomonas population, which here we named P. fuscovaginae-like, represents a perfect opportunity to understand common genomic features that can explain the infection mechanism in rice. We showed that the novel population is indeed closely related to P. fuscovaginae. A comparative genomics approach on eight rice-infecting Pseudomonas revealed heterogeneous genomes and a high number of strain-specific genes. The genomes of P. fuscovaginae-like harbor four secretion systems (Type I, II, III, and VI) and other important pathogenicity machinery that could probably facilitate rice colonization. We identified 123 core secreted proteins, most of which have strong signatures of positive selection suggesting functional adaptation. Transcript accumulation of putative pathogenicity-related genes during rice colonization revealed a concerted virulence mechanism. The study suggests that rice-infecting Pseudomonas causing sheath brown rot are intrinsically diverse and maintain a variable set of metabolic capabilities as a potential strategy to occupy a range of environments.

High-resolution image reconstruction from multiple low-resolution images.

In this paper, we demonstrate a digital signal processing (DSP) algorithm for improving spatial resolution of images captured by CMOS cameras. The basic approach is to reconstruct a high resolution (HR) image from a shift-related low resolution (LR) image sequence. The aliasing relationship of Fourier transforms between discrete and continuous images in the frequency domain is used for mapping LR images to a HR image. The method of projection onto convex sets (POCS) is applied to trace the best estimate of pixel matching from the LR images to the reconstructed HR image. Computer simulations and preliminary experimental results have shown that the algorithm works effectively on the application of post-image-captured processing for CMOS cameras. It can also be applied to HR digital image reconstruction, where shift information of the LR image sequence is known.