Matching Items (321)
Filtering by

Clear all filters

151716-Thumbnail Image.png
Description
The rapid escalation of technology and the widespread emergence of modern technological equipments have resulted in the generation of humongous amounts of digital data (in the form of images, videos and text). This has expanded the possibility of solving real world problems using computational learning frameworks. However, while gathering a

The rapid escalation of technology and the widespread emergence of modern technological equipments have resulted in the generation of humongous amounts of digital data (in the form of images, videos and text). This has expanded the possibility of solving real world problems using computational learning frameworks. However, while gathering a large amount of data is cheap and easy, annotating them with class labels is an expensive process in terms of time, labor and human expertise. This has paved the way for research in the field of active learning. Such algorithms automatically select the salient and exemplar instances from large quantities of unlabeled data and are effective in reducing human labeling effort in inducing classification models. To utilize the possible presence of multiple labeling agents, there have been attempts towards a batch mode form of active learning, where a batch of data instances is selected simultaneously for manual annotation. This dissertation is aimed at the development of novel batch mode active learning algorithms to reduce manual effort in training classification models in real world multimedia pattern recognition applications. Four major contributions are proposed in this work: $(i)$ a framework for dynamic batch mode active learning, where the batch size and the specific data instances to be queried are selected adaptively through a single formulation, based on the complexity of the data stream in question, $(ii)$ a batch mode active learning strategy for fuzzy label classification problems, where there is an inherent imprecision and vagueness in the class label definitions, $(iii)$ batch mode active learning algorithms based on convex relaxations of an NP-hard integer quadratic programming (IQP) problem, with guaranteed bounds on the solution quality and $(iv)$ an active matrix completion algorithm and its application to solve several variants of the active learning problem (transductive active learning, multi-label active learning, active feature acquisition and active learning for regression). These contributions are validated on the face recognition and facial expression recognition problems (which are commonly encountered in real world applications like robotics, security and assistive technology for the blind and the visually impaired) and also on collaborative filtering applications like movie recommendation.
ContributorsChakraborty, Shayok (Author) / Panchanathan, Sethuraman (Thesis advisor) / Balasubramanian, Vineeth N. (Committee member) / Li, Baoxin (Committee member) / Mittelmann, Hans (Committee member) / Ye, Jieping (Committee member) / Arizona State University (Publisher)
Created2013
152099-Thumbnail Image.png
Description
The body is capable of regulating hunger in several ways. Some of these hunger regulation methods are innate, such as genetics, and some, such as the responses to stress and to the smell of food, are innate but can be affected by body conditions such as BMI and physical activity.

The body is capable of regulating hunger in several ways. Some of these hunger regulation methods are innate, such as genetics, and some, such as the responses to stress and to the smell of food, are innate but can be affected by body conditions such as BMI and physical activity. Further, some hunger regulation methods stem from learned behaviors originating from cultural pressures or parenting styles. These latter regulation methods for hunger can be grouped into the categories: emotion, environment, and physical. The factors that regulate hunger can also influence the incidence of disordered eating, such as eating in the absence of hunger (EAH). Eating in the absence of hunger can occur in one of two scenarios, continuous EAH or beginning EAH. College students are at a particularly high risk for EAH and weight gain due to stress, social pressures, and the constant availability of energy dense and nutrient poor food options. The purpose of this study is to validate a modified EAH-C survey in college students and to discover which of the three latent factors (emotion, environment, physical) best predicts continual and beginning EAH. To do so, a modified EAH-C survey, with additional demographic components, was administered to students at a major southwest university. This survey contained two questions, one each for continuing and beginning EAH, regarding 14 factors related to emotional, physical, or environmental reasons that may trigger EAH. The results from this study revealed that the continual and beginning EAH surveys displayed good internal consistency reliability. We found that for beginning and continuing EAH, although emotion is the strongest predictor of EAH, all three latent factors are significant predictors of EAH. In addition, we found that environmental factors had the greatest influence on an individual's likelihood to continue to eat in the absence of hunger. Due to statistical abnormalities and differing numbers of factors in each category, we were unable to determine which of the three factors exerted the greatest influence on an individual's likelihood to begin eating in the absence of hunger. These results can be utilized to develop educational tools aimed at reducing EAH in college students, and ultimately reducing the likelihood for unhealthy weight gain and health complications related to obesity.
ContributorsGoett, Taylor (Author) / Johnston, Carol (Thesis advisor) / Lee, Chong (Committee member) / Lespron, Christy (Committee member) / Arizona State University (Publisher)
Created2013
151861-Thumbnail Image.png
Description
In October, 2009, participants of the Arizona Special Supplemental Nutrition Program for Women, Infants and Children (WIC) began receiving monthly Cash Value Vouchers (CVV) worth between six and 10 dollars towards the purchase of fresh fruits and vegetables. Data from the Arizona Department of Health Services (ADHS) showed CVV redemption

In October, 2009, participants of the Arizona Special Supplemental Nutrition Program for Women, Infants and Children (WIC) began receiving monthly Cash Value Vouchers (CVV) worth between six and 10 dollars towards the purchase of fresh fruits and vegetables. Data from the Arizona Department of Health Services (ADHS) showed CVV redemption rates in the first two years of the program were lower than the national average of 77% redemption. In response, the ADHS WIC Food List was expanded to also include canned and frozen fruits and vegetables. More recent data from ADHS suggest that redemption rates are improving, but variably exist among different WIC sub-populations. The purpose of this project was to identify themes related to the ease or difficulty of WIC CVV use amongst different categories of low-redeeming WIC participants. A total of 8 focus groups were conducted, four at a clinic in each of two Valley cities: Surprise and Mesa. Each of the four focus groups comprised one of four targeted WIC participant categories: pregnant, postpartum, breastfeeding, and children with participation ranging from 3-9 participants per group. Using the general inductive approach, recordings of the focus groups were transcribed, hand-coded and uploaded into qualitative analysis software resulting in four emergent themes including: interactions and shopping strategies, maximizing WIC value, redemption issues, and effect of rule change. Researchers identified twelve different subthemes related to the emergent theme of interactions and strategies to improve their experience, including economic considerations during redemption. Barriers related to interactions existed that made their purchase difficult, most notably anger from the cashier and other shoppers. However, participants made use of a number of strategies to facilitate WIC purchases or extract more value out of WIC benefits, such as pooling their CVV. Finally, it appears that the fruit and vegetable rule change was well received by those who were aware of the change. These data suggest a number of important avenues for future research, including verifying these themes are important within a larger, representative sample of Arizona WIC participants, and exploring strategies to minimize barriers identified by participants, such as use of electronic benefits transfer-style cards (EBT).
ContributorsBertmann, Farryl M. W (Author) / Wharton, Christopher (Christopher Mack), 1977- (Thesis advisor) / Ohri-Vachaspati, Punam (Committee member) / Johnston, Carol (Committee member) / Hampl, Jeffrey (Committee member) / Dixit-Joshi, Sujata (Committee member) / Barroso, Cristina (Committee member) / Arizona State University (Publisher)
Created2013
151926-Thumbnail Image.png
Description
In recent years, machine learning and data mining technologies have received growing attention in several areas such as recommendation systems, natural language processing, speech and handwriting recognition, image processing and biomedical domain. Many of these applications which deal with physiological and biomedical data require person specific or person adaptive systems.

In recent years, machine learning and data mining technologies have received growing attention in several areas such as recommendation systems, natural language processing, speech and handwriting recognition, image processing and biomedical domain. Many of these applications which deal with physiological and biomedical data require person specific or person adaptive systems. The greatest challenge in developing such systems is the subject-dependent data variations or subject-based variability in physiological and biomedical data, which leads to difference in data distributions making the task of modeling these data, using traditional machine learning algorithms, complex and challenging. As a result, despite the wide application of machine learning, efficient deployment of its principles to model real-world data is still a challenge. This dissertation addresses the problem of subject based variability in physiological and biomedical data and proposes person adaptive prediction models based on novel transfer and active learning algorithms, an emerging field in machine learning. One of the significant contributions of this dissertation is a person adaptive method, for early detection of muscle fatigue using Surface Electromyogram signals, based on a new multi-source transfer learning algorithm. This dissertation also proposes a subject-independent algorithm for grading the progression of muscle fatigue from 0 to 1 level in a test subject, during isometric or dynamic contractions, at real-time. Besides subject based variability, biomedical image data also varies due to variations in their imaging techniques, leading to distribution differences between the image databases. Hence a classifier learned on one database may perform poorly on the other database. Another significant contribution of this dissertation has been the design and development of an efficient biomedical image data annotation framework, based on a novel combination of transfer learning and a new batch-mode active learning method, capable of addressing the distribution differences across databases. The methodologies developed in this dissertation are relevant and applicable to a large set of computing problems where there is a high variation of data between subjects or sources, such as face detection, pose detection and speech recognition. From a broader perspective, these frameworks can be viewed as a first step towards design of automated adaptive systems for real world data.
ContributorsChattopadhyay, Rita (Author) / Panchanathan, Sethuraman (Thesis advisor) / Ye, Jieping (Thesis advisor) / Li, Baoxin (Committee member) / Santello, Marco (Committee member) / Arizona State University (Publisher)
Created2013
151963-Thumbnail Image.png
Description
Currently, to interact with computer based systems one needs to learn the specific interface language of that system. In most cases, interaction would be much easier if it could be done in natural language. For that, we will need a module which understands natural language and automatically translates it to

Currently, to interact with computer based systems one needs to learn the specific interface language of that system. In most cases, interaction would be much easier if it could be done in natural language. For that, we will need a module which understands natural language and automatically translates it to the interface language of the system. NL2KR (Natural language to knowledge representation) v.1 system is a prototype of such a system. It is a learning based system that learns new meanings of words in terms of lambda-calculus formulas given an initial lexicon of some words and their meanings and a training corpus of sentences with their translations. As a part of this thesis, we take the prototype NL2KR v.1 system and enhance various components of it to make it usable for somewhat substantial and useful interface languages. We revamped the lexicon learning components, Inverse-lambda and Generalization modules, and redesigned the lexicon learning algorithm which uses these components to learn new meanings of words. Similarly, we re-developed an inbuilt parser of the system in Answer Set Programming (ASP) and also integrated external parser with the system. Apart from this, we added some new rich features like various system configurations and memory cache in the learning component of the NL2KR system. These enhancements helped in learning more meanings of the words, boosted performance of the system by reducing the computation time by a factor of 8 and improved the usability of the system. We evaluated the NL2KR system on iRODS domain. iRODS is a rule-oriented data system, which helps in managing large set of computer files using policies. This system provides a Rule-Oriented interface langauge whose syntactic structure is like any procedural programming language (eg. C). However, direct translation of natural language (NL) to this interface language is difficult. So, for automatic translation of NL to this language, we define a simple intermediate Policy Declarative Language (IPDL) to represent the knowledge in the policies, which then can be directly translated to iRODS rules. We develop a corpus of 100 policy statements and manually translate them to IPDL langauge. This corpus is then used for the evaluation of NL2KR system. We performed 10 fold cross validation on the system. Furthermore, using this corpus, we illustrate how different components of our NL2KR system work.
ContributorsKumbhare, Kanchan Ravishankar (Author) / Baral, Chitta (Thesis advisor) / Ye, Jieping (Committee member) / Li, Baoxin (Committee member) / Arizona State University (Publisher)
Created2013
152003-Thumbnail Image.png
Description
We solve the problem of activity verification in the context of sustainability. Activity verification is the process of proving the user assertions pertaining to a certain activity performed by the user. Our motivation lies in incentivizing the user for engaging in sustainable activities like taking public transport or recycling. Such

We solve the problem of activity verification in the context of sustainability. Activity verification is the process of proving the user assertions pertaining to a certain activity performed by the user. Our motivation lies in incentivizing the user for engaging in sustainable activities like taking public transport or recycling. Such incentivization schemes require the system to verify the claim made by the user. The system verifies these claims by analyzing the supporting evidence captured by the user while performing the activity. The proliferation of portable smart-phones in the past few years has provided us with a ubiquitous and relatively cheap platform, having multiple sensors like accelerometer, gyroscope, microphone etc. to capture this evidence data in-situ. In this research, we investigate the supervised and semi-supervised learning techniques for activity verification. Both these techniques make use the data set constructed using the evidence submitted by the user. Supervised learning makes use of annotated evidence data to build a function to predict the class labels of the unlabeled data points. The evidence data captured can be either unimodal or multimodal in nature. We use the accelerometer data as evidence for transportation mode verification and image data as evidence for recycling verification. After training the system, we achieve maximum accuracy of 94% when classifying the transport mode and 81% when detecting recycle activity. In the case of recycle verification, we could improve the classification accuracy by asking the user for more evidence. We present some techniques to ask the user for the next best piece of evidence that maximizes the probability of classification. Using these techniques for detecting recycle activity, the accuracy increases to 93%. The major disadvantage of using supervised models is that it requires extensive annotated training data, which expensive to collect. Due to the limited training data, we look at the graph based inductive semi-supervised learning methods to propagate the labels among the unlabeled samples. In the semi-supervised approach, we represent each instance in the data set as a node in the graph. Since it is a complete graph, edges interconnect these nodes, with each edge having some weight representing the similarity between the points. We propagate the labels in this graph, based on the proximity of the data points to the labeled nodes. We estimate the performance of these algorithms by measuring how close the probability distribution of the data after label propagation is to the probability distribution of the ground truth data. Since labeling has a cost associated with it, in this thesis we propose two algorithms that help us in selecting minimum number of labeled points to propagate the labels accurately. Our proposed algorithm achieves a maximum of 73% increase in performance when compared to the baseline algorithm.
ContributorsDesai, Vaishnav (Author) / Sundaram, Hari (Thesis advisor) / Li, Baoxin (Thesis advisor) / Turaga, Pavan (Committee member) / Arizona State University (Publisher)
Created2013
151655-Thumbnail Image.png
Description
There are several visual dimensions of food that can affect food intake, example portion size, color, and variety. This dissertation elucidates the effect of number of pieces of food on preference and amount of food consumed in humans and motivation for food in animals. Chapter 2 Experiment 1 showed that

There are several visual dimensions of food that can affect food intake, example portion size, color, and variety. This dissertation elucidates the effect of number of pieces of food on preference and amount of food consumed in humans and motivation for food in animals. Chapter 2 Experiment 1 showed that rats preferred and also ran faster for multiple pieces (30, 10 mg pellets) than an equicaloric, single piece of food (300 mg) showing that multiple pieces of food are more rewarding than a single piece. Chapter 2 Experiment 2 showed that rats preferred a 30-pellet food portion clustered together rather than scattered. Preference and motivation for clustered food pieces may be interpreted based on the optimal foraging theory that animals prefer foods that can maximize energy gain and minimize the risk of predation. Chapter 3 Experiment 1 showed that college students preferred and ate less of a multiple-piece than a single-piece portion and also ate less in a test meal following the multiple-piece than single-piece portion. Chapter 3 Experiment 2 replicated the results in Experiment 1 and used a bagel instead of chicken. Chapter 4 showed that college students given a five-piece chicken portion scattered on a plate ate less in a meal and in a subsequent test meal than those given the same portion clustered together. This is consistent with the hypothesis that multiple pieces of food may appear like more food because they take up a larger surface area than a single-piece portion. All together, these studies show that number and surface area occupied by food pieces are important visual cues determining food choice in animals and both food choice and intake in humans.
ContributorsBajaj, Devina (Author) / Phillips, Elizabeth D. (Thesis advisor) / Cohen, Adam (Committee member) / Johnston, Carol (Committee member) / Bimonte-Nelson, Heather A. (Committee member) / Arizona State University (Publisher)
Created2013
151284-Thumbnail Image.png
Description
Dietary protein is known to increase postprandial thermogenesis more so than carbohydrates or fats, probably related to the fact that amino acids have no immediate form of storage in the body and can become toxic if not readily incorporated into body tissues or excreted. It is also well documented that

Dietary protein is known to increase postprandial thermogenesis more so than carbohydrates or fats, probably related to the fact that amino acids have no immediate form of storage in the body and can become toxic if not readily incorporated into body tissues or excreted. It is also well documented that subjects report greater satiety on high- versus low-protein diets and that subject compliance tends to be greater on high-protein diets, thus contributing to their popularity. What is not as well known is how a high-protein diet affects resting metabolic rate over time, and what is even less well known is if resting metabolic rate changes significantly when a person consuming an omnivorous diet suddenly adopts a vegetarian one. This pilot study sought to determine whether subjects adopting a vegetarian diet would report decreased satiety or demonstrate a decreased metabolic rate due to a change in protein intake and possible increase in carbohydrates. Further, this study sought to validate a new device called the SenseWear Armband (SWA) to determine if it might be sensitive enough to detect subtle changes in metabolic rate related to diet. Subjects were tested twice on all variables, at baseline and post-test. Independent and related samples tests revealed no significant differences between or within groups for any variable at any time point in the study. The SWA had a strong positive correlation to the Oxycon Mobile metabolic cart but due to a lack of change in metabolic rate, its sensitivity was undetermined. These data do not support the theory that adopting a vegetarian diet results in a long-term change in metabolic rate.
ContributorsMoore, Amy (Author) / Johnston, Carol (Thesis advisor) / Appel, Christy (Thesis advisor) / Gaesser, Glenn (Committee member) / Arizona State University (Publisher)
Created2012
151503-Thumbnail Image.png
Description
Objective: Vinegar consumption studies have demonstrated possible therapeutic effects in reducing HbA1c and postprandial glycemia. The purpose of the study was to closely examine the effects of a commercial vinegar drink on daily fluctuations in fasting glucose concentrations and postprandial glycemia, and on HbA1c, in individuals at risk for Type

Objective: Vinegar consumption studies have demonstrated possible therapeutic effects in reducing HbA1c and postprandial glycemia. The purpose of the study was to closely examine the effects of a commercial vinegar drink on daily fluctuations in fasting glucose concentrations and postprandial glycemia, and on HbA1c, in individuals at risk for Type 2 Diabetes Mellitus (T2D). Design: Thirteen women and one man (21-62 y; mean, 46.0±3.9 y) participated in this 12-week parallel-arm trial. Participants were recruited from a campus community and were healthy and not diabetic by self-report. Participants were not prescribed oral hypoglycemic medications or insulin; other medications were allowed if use was stable for > 3 months. Subjects were randomized to one of two groups: VIN (8 ounces vinegar drink providing 1.5 g acetic acid) or CON (1 vinegar pill providing 0.04 g acetic acid). Treatments were taken twice daily immediately prior to the lunch and dinner meals. Venous blood samples were drawn at trial weeks 0 and 12 to measure insulin, fasting glucose, and HbA1c. Subjects recorded fasting glucose and 2-h postprandial glycemia concentrations daily using a glucometer. Results: The VIN group showed significant reductions in fasting capillary blood glucose concentrations (p=0.05) that were immediate and sustained throughout the duration of the study. The VIN group had reductions in 2-h postprandial glucose (mean change of −7.6±6.8 mg/dL over the 12-week trial), but this value was not significantly different than that for the CON group (mean change of 3.3±5.3 mg/dL over the 12-week trial, p=0.232). HbA1c did not significantly change (p=0.702), but the reduction in HbA1c in the VIN group, −0.14±0.1%, may have physiological relevance. Conclusions: Significant reductions in HbA1c were not observed after daily consumption of a vinegar drink containing 1.5 g acetic acid in non-diabetic individuals. However, the vinegar drink did significantly reduce fasting capillary blood glucose concentrations in these individuals as compared to a vinegar pill containing 0.04 g acetic acid. These results support a therapeutic effect for vinegar in T2D prevention and progression, specifically in high-risk populations.
ContributorsQuagliano, Samantha (Author) / Johnston, Carol (Thesis advisor) / Appel, Christy (Committee member) / Dixon, Kathleen (Committee member) / Arizona State University (Publisher)
Created2013
151504-Thumbnail Image.png
Description
Objective: The purpose of this randomized parallel arm trial was to demonstrate the effects of daily fish oil supplementation (600mg per day for eight weeks) on body composition and body mass in young healthy women, aged 18-38, at a large southwestern university. Design: 26 non-obese (mean BMI 23.7±0.6 kg/m2), healthy

Objective: The purpose of this randomized parallel arm trial was to demonstrate the effects of daily fish oil supplementation (600mg per day for eight weeks) on body composition and body mass in young healthy women, aged 18-38, at a large southwestern university. Design: 26 non-obese (mean BMI 23.7±0.6 kg/m2), healthy women (18-38y; mean, 23.5±1.1 y) from a southwestern Arizona university campus community completed the study. Subjects were healthy, non-smokers, consuming less than 3.5 oz of fish per week according to self-report. Participants were randomized to one of two groups: FISH (600 mg omega-3 fatty acids provided in one gel capsule per day), or CON (1000 mg coconut oil placebo provided in one gel capsule per day). Body weight, BMI, and percent body fat were measured using a stadiometer and bioelectrical impedance scale at the screening visit and intervention weeks 1, 4, and 8. 24-hour dietary recalls were also performed at weeks 1 and 8. Results: 8 weeks of omega-3 fatty acid supplementation did not significantly alter body weight (p=0.830), BMI (p=1.00), or body fat percentage (p=0.600) as compared to placebo. Although not statistically significant, 24-hour dietary recalls performed at the beginning and end of the intervention revealed a trend towards increased caloric intake in the FISH group and decreased caloric intake in the CON group throughout the course of the study (p=0.069). If maintained, this difference in caloric intake could have physiological relevance. Conclusions: Omega-3 fatty acids do not significantly alter body weight or body composition in healthy young females. These findings do not refute the current recommendations for Americans to consume at least 8 oz of omega-3-rich seafood per week, supplying 250 mg EPA and DHA per day. More research is needed to investigate the potential for omega-3 fatty acids to modulate daily caloric intake.
ContributorsTeran, Bianca (Author) / Johnston, Carol (Thesis advisor) / Johnson, Melinda (Committee member) / Ohri-Vachaspati, Punam (Committee member) / Arizona State University (Publisher)
Created2013