Graph neural networks (GNN) are vital for analyzing real-world problems (e.g., network analysis, drug interaction, electronic design automation, e-commerce) that use graph models. However, efficient GNN acceleration faces with multiple challenges related to high and variable sparsity of input feature vectors, power-law degree distribution in the…

All communication systems rely on sets of rules that must generally be followed for successful communication to be possible; in the case of human language, these rules are given by the grammar. However, despite these rules, there is considerable variation that has been noted across the grammar. In this dissertation, I explore one understudied…

The demographics of our nation's public schools continue to diversify. Less than 50% of our nation's public school students are white, however, less than 25% percent of superintendents identify as People of Color. To that end, retention of diverse leaders specifically, Black Women across all sectors is in decline. The mission of The Surge…

China, classified as a middle-income country by the World Bank, faces a scarcity of resources in providing autism interventions, despite a comparable autism prevalence in the Western world. Most autism interventions in China are imported from high-income countries, highlighting the need to examine whether the focus of the interventions from mixed…

Teaching reading has been an ongoing challenge for many educators across the United States. Numerous studies suggest that teachers feel they were not adequately prepared to teach reading and have low efficacy in delivering literacy instruction. Teaching reading requires educators to be knowledgeable about the alphabetic principle, decoding new…

Categorical data become increasingly ubiquitous in the modern big data era. In this dissertation, we propose novel statistical learning and inference methods for large-scale categorical data, focusing on latent variable models and their applications to psychometrics. In psychometric assessments, the subjects' underlying aptitude often cannot be…

This dissertation presents two essays on econometric methods for program evaluation and policy choice. In Chapter 1, I develop a statistically optimal way of using data to make policy decisions when the performance of counterfactual policies is only partially identified. Specifically, I consider a class of statistical decision problems in which…

Diversity in the science, technology, engineering, and mathematics (STEM) workforce is in continual demand among underrepresented minorities and women. Of particular concern is the disproportionality low rates of Black women in STEM in relationship to their science and engineering (S&E) degree attainment and overall representation in the…

Graph representations underlie many modern computer applications, capturing the structure of such diverse networks as the Internet, personal associations, roads, sensors, and metabolic pathways. While the static structure of graphs is a well-explored field, a new emphasis is being placed on understanding and representing the way these networks…

School psychologists use intelligence and academic achievement test scores to understand individual differences in abilities that may be directed towards learning and the learning that has already occurred. To make ethical and effective decisions regarding individualized educational programs and interventions, test users must understand the…

With close to 50 million children attending 98,000 public schools in the United States, the public school classroom continues to be an essential locus of change that, if improved at scale, could increase educational and life outcomes of many young people. Improved student learning requires an improved instructional core and school leaders are…

The World Wide Web enables an easy, instant access to a huge quantity of information. Over the last few decades, a number of improvements have been achieved that helped the web reach its current state. However, the current Internet links documents together without understanding them, and thus, makes the content of web only human-readable rather…

Statewide accountability programs are incorporating academic growth estimates for general assessments. This transition focuses attention on modeling growth for students with significant cognitive disabilities (SWSCD) who take alternate assessments based on alternate achievement standards (AA-AAS), as most states attempt to structure their AA-AAS…

Two problems at two different abstraction levels of computational biology are studied. At the molecular level, efficient pattern matching algorithms in DNA sequences are presented. For gene order data, an efficient data structure is presented capable of storing all gene re-orderings in a systematic manner. A common characteristic of presented…

This study aims to evaluate a multidimensional latent trait model to determine how well the model works in various empirical contexts. Contrary to the assumption of these latent trait models that the traits are normally distributed, situations in which the latent trait is not shaped with a normal distribution may occur (Sass et al, 2008; Woods…