On Leveraging Representation Learning Techniques for Data Analytics in Biomedical Informatics
AuthorCao, Xi Hang
Committee memberVucetic, Slobodan
DepartmentComputer and Information Science
Permanent link to this recordhttp://hdl.handle.net/20.500.12613/903
MetadataShow full item record
AbstractRepresentation Learning is ubiquitous in state-of-the-art machine learning workflow, including data exploration/visualization, data preprocessing, data model learning, and model interpretations. However, the majority of the newly proposed Representation Learning methods are more suitable for problems with a large amount of data. Applying these methods to problems with a limited amount of data may lead to unsatisfactory performance. Therefore, there is a need for developing Representation Learning methods which are tailored for problems with ``small data", such as, clinical and biomedical data analytics. In this dissertation, we describe our studies of tackling the challenging clinical and biomedical data analytics problem from four perspectives: data preprocessing, temporal data representation learning, output representation learning, and joint input-output representation learning. Data scaling is an important component in data preprocessing. The objective in data scaling is to scale/transform the raw features into reasonable ranges such that each feature of an instance will be equally exploited by the machine learning model. For example, in a credit flaw detection task, a machine learning model may utilize a person's credit score and annual income as features, but because the ranges of these two features are different, a machine learning model may consider one more heavily than another. In this dissertation, I thoroughly introduce the problem in data scaling and describe an approach for data scaling which can intrinsically handle the outlier problem and lead to better model prediction performance. Learning new representations for data in the unstandardized form is a common task in data analytics and data science applications. Usually, data come in a tubular form, namely, the data is represented by a table in which each row is a feature (row) vector of an instance. However, it is also common that the data are not in this form; for example, texts, images, and video/audio records. In this dissertation, I describe the challenge of analyzing imperfect multivariate time series data in healthcare and biomedical research and show that the proposed method can learn a powerful representation to encounter various imperfections and lead to an improvement of prediction performance. Learning output representations is a new aspect of Representation Learning, and its applications have shown promising results in complex tasks, including computer vision and recommendation systems. The main objective of an output representation algorithm is to explore the relationship among the target variables, such that a prediction model can efficiently exploit the similarities and potentially improve prediction performance. In this dissertation, I describe a learning framework which incorporates output representation learning to time-to-event estimation. Particularly, the approach learns the model parameters and time vectors simultaneously. Experimental results do not only show the effectiveness of this approach but also show the interpretability of this approach from the visualizations of the time vectors in 2-D space. Learning the input (feature) representation, output representation, and predictive modeling are closely related to each other. Therefore, it is a very natural extension of the state-of-the-art by considering them together in a joint framework. In this dissertation, I describe a large-margin ranking-based learning framework for time-to-event estimation with joint input embedding learning, output embedding learning, and model parameter learning. In the framework, I cast the functional learning problem to a kernel learning problem, and by adopting the theories in Multiple Kernel Learning, I propose an efficient optimization algorithm. Empirical results also show its effectiveness on several benchmark datasets.
ADA complianceFor Americans with Disabilities Act (ADA) accommodation, including help with reading this content, please contact firstname.lastname@example.org
Showing items related by title, author, creator and subject.
Is international service-learning win-win? A case study of an engineering partnershipHorvat, Erin McNamara; Kaplan, Avi; Keith, Novella Zett; Hartman, Eric; Brandt, Carol B. (Temple University. Libraries, 2016)Given the tormented history of development projects around the globe and the fact that global service learning and engineering-for-development often engages students in development interventions, it is critically important to explore the impact of global service learning projects and partnerships not only on student learning, which has received ample attention, but also on the international host communities. Although there is increasing research on student outcomes of participation in service-learning, there is a lack of research focusing on the outcomes related to the community where the service takes place. Research focused on the impact on communities should include the wide range of perspectives that compose the community- participants, organization leaders, residents, and others. This study responds to this need by exploring the community participants’ perspectives in Waslala, Nicaragua about the projects and partnership with Villanova University’s College of Engineering. The two research questions explore the community participants’ perspectives about: (1) outcomes of the projects or partnership, and (2) educational goals. Community participants’ perspectives call for a broader conceptualization of what counts as outcomes and highlight the importance of participation in all phases of the research process. Supporting existing scholarship about host community motivations to serve as co-educators, my findings describe what the community participants want to teach university students. Community participants’ perspectives describe a desired shift in students and a move towards critical global citizenship education. This study highlights the utility of Fraser’s (2009) theory of social justice and Andreotti’s (2006) framework for critical global citizenship education as useful tools to analyze and understand GSL partnerships.
Motivational trajectories of successful foreign language learners: Six biographical case studiesBeglar, David J.; Nation, Paul; Sawyer, Mark; Takeuchi, Osamu; Irie, Kay (Temple University. Libraries, 2011)This study concerns foreign language learners' motivational changes over a long period of time; it is an investigation of the learning histories of six learners who have achieved high proficiency in English. Unlike a large body of conventional foreign language learning motivational research, which has primarily been conducted using quantitative methodologies, this study employs two non-conventional approaches, a combination of learners' biographies and case study research. The primary purpose of the study is to holistically explore successful English learners' motivational trajectories and their learning histories in the Japanese context. To this end, foreign language learning motivation is conceptualized and illustrated as a dynamically changing construct that plays an important role in the process of foreign language learning. In the literature review, longitudinal studies concerning foreign language learning motivation and autobiographical studies and case studies that are relevant to this study are examined. The central research question is what motivational trajectories and learning histories these highly proficient learners have had, and how these learners have sustained their learning motivation over time and eventually achieved high proficiency while in an EFL (English as a foreign language) environment. The participants are six Japanese adults who have achieved high levels of English proficiency and who use English in their jobs. The design used in this case study involves both holistic and specifically focused analyses, by which each participant's learning history is collected through individual interviews. The author reports each participant's learning history, and the initial proposition concerning motivational change and salient motivational sources found in the participants' learning histories are collectively analyzed and discussed. Exploring the data concerning how the participants have maintained foreign language learning motivation resulting in the idea that sustained motivation is not always present in successful foreign language learning and that the key to success involves a cognitive change from a state in which motivation is present to one in which a more intentional psychological force, commitment to learning, develops. Based on this thought, a model illustrating the key to success in foreign language learning in the EFL context is presented. The results provide new, engaging, and important information to people who are seriously involved in foreign language learning in EFL contexts, where the majority of learners fail to attain high levels of foreign language proficiency after receiving years of formal education.
MANAGING TRANSFORMATION: HOW DO UNDERGRADUATE STUDENTS EXPERIENCE THE PROCESS OF REENTRY AFTER INTERNATIONAL SERVICE LEARNING?Gross, Steven Jay; Belliveau, Cynthia S.; Davis, James Earl (Temple University. Libraries, 2017)As calls for institutions of higher education to educate globally competent citizens have intensified over the last two decades, the field of international service learning (ISL) has responded resoundingly. ISL programs have been implemented at many institutions and there have been countless studies that demonstrate the great power for student learning and growth inherent in this exciting educational tool. In more recent years, experts have moved away from studying the student experience and have, instead, questioned the power relationships inherent in service learning programs abroad; related studies have made use of newer critical theories and community development philosophies, which have advanced the field tremendously. However, to date, the re-entry period has still been largely overlooked, and there has been a noticeable lack of studies that apply student development theories to the ISL experience. This study explores how a diverse group of students from a large, public, four-year institution on the West Coast experience transformational learning during an ISL program in Thailand and how they make sense of their experience upon their return to the U.S. and in the months afterwards. By utilizing a case study design and implementing qualitative methods, this study provides significant evidence that well-designed ISL programs can trigger transformational learning in a variety of ways and that the re-entry period is a significant time of learning and growth for students. Further, the findings importantly show that by creating strategic opportunities for students to learn and connect with others on the program, both during and after the ISL experience, students are better able to navigate the changes in themselves after returning home. Finally, the experiences of the students indicate that the processes associated with transformational learning continue well on after the in-country experience, highlighting the significant need to provide support and guidance for students during this time.