Leveraging Multiple Features for Image Retrieval and Matching

The goal of image retrieval and matching is to find and locate object instances in images from a large-scale image database. While visual features are abundant, how to combine them to improve performance by individual features remains a challenging task. In this work, we focus on leveraging multiple features for accurate and efficient image retrieval and matching. We first propose two graph-based approaches to rerank initially retrieved images for generic image retrieval. In the graph, vertices are images while edges are similarities between image pairs. Our first approach employs a mixture Markov model based on a random walk model on multiple graphs to fuse graphs. We introduce a probabilistic model to compute the importance of each feature for graph fusion under a naive Bayesian formulation, which requires statistics of similarities from a manually labeled dataset containing irrelevant images. To reduce human labeling, we further propose a fully unsupervised reranking algorithm based on a submodular objective function that can be efficiently optimized by greedy algorithm. By maximizing an information gain term over the graph, our submodular function favors a subset of database images that are similar to query images and resemble each other. The function also exploits the rank relationships of images from multiple ranked lists obtained by different features. We then study a more well-defined application, person re-identification, where the database contains labeled images of human bodies captured by multiple cameras. Re-identifications from multiple cameras are regarded as related tasks to exploit shared information. We apply a novel multi-task learning algorithm using both low level features and attributes. A low rank attribute embedding is joint learned within the multi-task learning formulation to embed original binary attributes to a continuous attribute space, where incorrect and incomplete attributes are rectified and recovered. To locate objects in images, we design an object detector based on object proposals and deep convolutional neural networks (CNN) in view of the emergence of deep networks. We improve a Fast RCNN framework and investigate two new strategies to detect objects accurately and efficiently: scale-dependent pooling (SDP) and cascaded rejection classifiers (CRC). The SDP improves detection accuracy by exploiting appropriate convolutional features depending on the scale of input object proposals. The CRC effectively utilizes convolutional features and greatly eliminates negative proposals in a cascaded manner, while maintaining a high recall for true objects. The two strategies together improve the detection accuracy and reduce the computational cost.

Optimal Budget-Constrained Sample Allocation for Selection Decisions with Multiple Uncertain Attributes

A decision-maker, when faced with a limited and fixed budget to collect data in support of a multiple attribute selection decision, must decide how many samples to observe from each alternative and attribute. This allocation decision is of particular importance when the information gained leads to uncertain estimates of the attribute values as with sample data collected from observations such as measurements, experimental evaluations, or simulation runs. For example, when the U.S. Department of Homeland Security must decide upon a radiation detection system to acquire, a number of performance attributes are of interest and must be measured in order to characterize each of the considered systems. We identified and evaluated several approaches to incorporate the uncertainty in the attribute value estimates into a normative model for a multiple attribute selection decision. Assuming an additive multiple attribute value model, we demonstrated the idea of propagating the attribute value uncertainty and describing the decision values for each alternative as probability distributions. These distributions were used to select an alternative. With the goal of maximizing the probability of correct selection we developed and evaluated, under several different sets of assumptions, procedures to allocate the fixed experimental budget across the multiple attributes and alternatives. Through a series of simulation studies, we compared the performance of these allocation procedures to the simple, but common, allocation procedure that distributed the sample budget equally across the alternatives and attributes. We found the allocation procedures that were developed based on the inclusion of decision-maker knowledge, such as knowledge of the decision model, outperformed those that neglected such information. Beginning with general knowledge of the attribute values provided by Bayesian prior distributions, and updating this knowledge with each observed sample, the sequential allocation procedure performed particularly well. These observations demonstrate that managing projects focused on a selection decision so that the decision modeling and the experimental planning are done jointly, rather than in isolation, can improve the overall selection results.

The relationship of early social, mental, and behavioral experiences with adult obesity and Alcohol Use Disorder

Stressful life events early in life, including symptoms of mental disorders or childhood maltreatment, may increase risk for worse mental and physical health outcomes in adulthood. The purpose of this dissertation was to examine the effects of childhood Attention Deficit Hyperactivity Disorder (ADHD) symptoms and maltreatment experience on two adult outcomes: obesity and alcohol use disorder (AUD). Mediational effects of adolescent characteristics were explored. This dissertation used Waves I, III, and IV of the National Longitudinal Study of Adolescent to Adult Health. In Paper 1 (Chapter 3), we investigated the association between multiple types of child maltreatment and adult objective (body mass index; BMI) and subjective (self-rated) obesity, as well as mediating effects by adolescent characteristics including depressive symptoms and BMI. Results showed that after adjusting for sex, race/ethnicity, and maternal education, physical maltreatment was moderately associated with adulthood obesity as measured by BMI and self-reported obesity, while sexual maltreatment was more strongly associated with the objective measure but not the subjective measure. The indirect effects of mediation of adolescent BMI and depressive symptoms were statistically significant. In Paper 2 (Chapter 4), the objective was to examine mediation by adolescent depressive symptoms, alcohol consumption, peer alcohol consumption, and delinquency in the relationship between ADHD symptoms and adult AUD. The indirect effects of mediation of adolescent delinquency, alcohol consumption, and peer alcohol consumption were statistically significant in single and multiple mediator models. In Paper 3 (Chapter 5), the objective was to assess the joint effects of maltreatment/neglect on adult AUD. After adjusting for sex, race/ethnicity, child maltreatment, and parental AUD, ADHD symptoms were significantly associated with increased odds of AUD. There was no strong evidence of multiplicative interaction by maltreatment. This association was stronger for males than females, although the interaction term was not statistically significant. This dissertation adds to the literature by examining relationships between several major public health problems: ADHD symptoms, childhood maltreatment, AUD, depressive symptoms, and obesity. This project has implications for understanding how early life stress increases risk for later physical and mental health problems, and identifying potential intervention targets for adolescents.

Probabilistic Models for Scalable Knowledge Graph Construction

In the past decade, systems that extract information from millions of Internet documents have become commonplace. Knowledge graphs -- structured knowledge bases that describe entities, their attributes and the relationships between them -- are a powerful tool for understanding and organizing this vast amount of information. However, a significant obstacle to knowledge graph construction is the unreliability of the extracted information, due to noise and ambiguity in the underlying data or errors made by the extraction system and the complexity of reasoning about the dependencies between these noisy extractions. My dissertation addresses these challenges by exploiting the interdependencies between facts to improve the quality of the knowledge graph in a scalable framework. I introduce a new approach called knowledge graph identification (KGI), which resolves the entities, attributes and relationships in the knowledge graph by incorporating uncertain extractions from multiple sources, entity co-references, and ontological constraints. I define a probability distribution over possible knowledge graphs and infer the most probable knowledge graph using a combination of probabilistic and logical reasoning. Such probabilistic models are frequently dismissed due to scalability concerns, but my implementation of KGI maintains tractable performance on large problems through the use of hinge-loss Markov random fields, which have a convex inference objective. This allows the inference of large knowledge graphs using 4M facts and 20M ground constraints in 2 hours. To further scale the solution, I develop a distributed approach to the KGI problem which runs in parallel across multiple machines, reducing inference time by 90%. Finally, I extend my model to the streaming setting, where a knowledge graph is continuously updated by incorporating newly extracted facts. I devise a general approach for approximately updating inference in convex probabilistic models, and quantify the approximation error by defining and bounding inference regret for online models. Together, my work retains the attractive features of probabilistic models while providing the scalability necessary for large-scale knowledge graph construction. These models have been applied on a number of real-world knowledge graph projects, including the NELL project at Carnegie Mellon and the Google Knowledge Graph.

Playing for Real: Designing Alternate Reality Games in Learning Contexts

Alternate Reality Game (ARG) represent a new genre of transmedia practice where players hunt for scattered clues, make sense of disparate information, and solve puzzles to advance an ever-evolving storyline. Players participate in ARGs using multiple communications technologies, ranging from print materials to mobile devices. However, many interaction design challenges must be addressed to weave these everyday communication tools together into an immersive, participatory experience. Transmedia design is not an everyday process. Designers must create and connect story bits across multiple media (video, audio, text) and multiple platforms (phones, computers, physical spaces). Furthermore, they must engage with players of varying skill levels. Few studies to-date have explored the design process of ARGs in learning contexts. Fewer still have focused on challenges involved in designing for youth (13-17 years old). In this study, I explore the process of designing ARGs as vehicles for promoting information literacy and participatory culture for adolescents (13-17 years old). Two ARG design scenarios, distinguished by target learning environment (formal and informal context) and target audience (adolescents), comprise the two cases that I examine. Through my analysis of these two design cases, I articulate several unique challenges faced by designers who create interactive, transmedia stories for – and with – youth. Drawing from these design challenges, I derive a repertoire of design strategies that future designers and researchers may use to create and implement ARGs for teens in learning contexts. In particular, I propose a narrative design framework that allows for the categorization of ARGs as storytelling constructs that lie along a continuum of participation and interaction. The framework can serve as an analytic tool for researchers and a guide for designers. In addition, I establish a framework of social roles that designers may employ to craft transmedia narratives before live launch and to promote and scaffold player participation after play begins. Overall, the contributions of my study include theoretical insights that may advance our understanding of narrative design and analysis as well as more practical design implications for designers and practitioners seeking to incorporate transmedia features into learning experiences that target youth.