Dissertation Defense Announcements

Federal legislation has mandated students with and without disabilities be prepared for college and careers (ESSA, 2015; IDEA, 2004). Students with high-incidence disabilities experience less success than their peers without disabilities (Newman et al., 2011). Initially, college and career readiness efforts lacked a focus on students with disabilities (e.g., Conley 2007, 2008), but recent efforts have increased the focus on students with disabilities (e.g., Morningstar et al., 2017). The predictors of post-school success appear to be a viable option to bridge both efforts. Students with high-incidence disabilities spend at least part of their day in general education classes (NCES, 2017), but general education teachers report wanting additional information to prepare students with high-incidence disabilities for college and careers (Kwiatek, 2017). General educators identified the predictors of post-school success as relevant, important, and feasible for implementation (Kwiatek et al., 2021). Coupling the alignment between secondary transition and college and career readiness, the predictors of post-school success appear to be an ideal option to provide general educators with professional development to prepare students with high-incidence disabilities for college and careers. The purpose of this dissertation was to examine the effects of an asynchronous online intervention (i.e., General Educators Now Embedding Research [for] Adult Life in Educational Design [GENERAL ED]) on general education teachers’ knowledge of research-based, in-school predictors of post-school success. Results indicated a functional relation between the asynchronous online intervention and increased knowledge of three predictors of post-school success. Effect sizes were large for increased knowledge of the predictors of post-school success. Additional measures included application; confidence; generalization; and social validity (i.e., feasibility evaluation, intervention rating scale). Finally, limitations, suggestions for future research, and implications for practice will be discussed.

The availability of OpenCL for FPGAs along with High-Level Synthesis tools have made FPGAs an attractive platform for realizing massively parallel compute-intensive applications. FPGAs with their customizable data-path, deep pipelining abilities and enhanced power efficiency features are the most viable solutions for programming and integrating them with heterogeneous platforms. Furthermore, OpenCL for FPGAs raises many challenges which require in-depth understanding to better utilize their enormous capabilities. In this work we identify, analyze and categorize the semantic differences between the OpenCL parallelism and the execution model on FPGAs. As an end result we propose a generic taxonomy for classifying FPGA parallelism potential.

At the same time, new design challenges continue to emerge for massive thread-level parallelism on FPGAs. One major execution bottleneck is the high number of memory stalls exposed to data-path which overshadows the benefits of data-path customization. We introduce a unique approach for hiding the memory stalls on FPGAs when running massively parallel applications and present a novel LLVM-based tool for decoupling memory access from computations. To enable systematic decoupling, we use the idea of kernel parallelism and implement a new parallelism granularity that breaks down kernels to separate data-path and memory-path (memory read/write) which work concurrently to overlap the computation of current threads with the memory access of future threads (memory pre-fetching at large scale).

We next move to the Xilinx based AWS cloud platform and conduct an exhaustive study on the scalability of OpenCL coarse-grain parallelism, Compute Unit(CU) replication on cloud FPGAs. In addition we present a generic template and a front-end design exploration tool to explore and identify the optimum CU number for a given application, while hiding the programming and exploration difficulties from programmers.

Special education continues to be the most litigated area of education with minimal requirements for the local education authority, typically an administrator, to have a fundamental understanding of special education law and practices. Additionally, federal, state, and local accountability measures indicate a significant achievement gap between students with disabilities and their peers. Given these concerns, there is a need for research on ways to improve administrator’s knowledge about special education as well as close the achievement gap between the two groups.

This qualitative study aims to explore perceptions about the role of an administrator for special education programming in order to identify specific areas of special education programming knowledge administrators need to be effective for special education. Interviews were conducted with three key stakeholder groups: special education teachers, general education teachers, and administrators, in order to ascertain converging and diverging perspectives about the role of an administrator for special education programming.

Findings in this study supported prior work around the lack of knowledge administrators had about special education law and practice. However, the findings went further to explore the skills needed to implement the knowledge administrators need in order to be effective for special education programming. Specifically, this study found that educational philosophy was an influential aspect to overall effectiveness of administrators for special education. Additionally, the ability to advocate (or champion) for the betterment of all stakeholders, and growth mindset were identified as key themes participants felt were necessary for an administrator to be effective for special education programming. These findings support the need for additional training that education administrators should receive to not only understand special education law, but also how to be an effective administrator for special education programming.

 KEY WORDS: Special education, School administrators and special education, Administrator preparation, Systems Thinking Theory, Administrator roles

Thermal stress has been identified as one of the major failure causes in the power module. It is generated from the mechanical strain by severely varying temperatures at different loci in the power module and the different coefficients of the thermal expansion of materials, where the varying temperatures result from the real-time power loss across the power converter. This thermal stress accelerates the degradation of semiconductor devices, downgrades the system quality and efficiency, and eventually causes catastrophic system breakdowns and extensive economic losses. Therefore, this research is dedicated to investigating both local control level methods and system level strategies to ameliorate the real-time power loss in order to reduce the thermal stress in the power module, thereby extend the component lifetime and enhance the system reliability. A finite-control-set model predictive control (FCS-MPC) is introduced and deductively investigated from the local control level. Its variable switching frequency property is derived through the geometry analysis on the voltage vector space. It realizes the switching frequency variation autonomously by the loading power. By taking advantage of this property, the power loss is leveled in the real-time operation by FCS-MPC, and a more mitigated thermal profile is acquired compared with the one by the conventional controller. Furthermore, a centralized thermal stress oriented dispatch (TSOD) system level strategy is proposed for multiple paralleled distributed energy resource systems, which helps to reduce the thermal stress in the power module of paralleled converters. It is thermal stress oriented and takes effect according to the real-time junction temperature variation, the health condition of the individual converter, and the system operation. Two local control level methods, the switching frequency variation and the reactive power injection, are imported separately as the dispatch algorithm to generate the expected power loss. Dealing with the varying mission profile, the more mitigated thermal profiles are achieved for all converters with the assistance of the proposed TSOD strategy.