Microscopy is a fundamental tool in scientific research, allowing the visualization ofstructures not visible to the naked eye. However, imaging thick tissue samples presents challenges due to light scattering from mismatched refractive indices. Tissue clearing techniques solve these issues by rendering tissues transparent, enhancing image quality and contrast-to-background ratios. Despite benefits, common methods can result in tissue deformation, long processing times, and fluorescence quenching. To address these issues, two advanced techniques have been developed: immunolabeling-enabled solvent-cleared organs (iDISCO) and the Fast Optical Clearing Method (FOCM). iDISCO is highly effective in large-scale tissue clearing, ideal for comprehensive studies, but uses hazardous chemicals and requires days to clear tissues. In contrast, FOCM offers a safer, more rapid alternative, achieving tissue transparency in as little as five minutes while maintaining vessel intensity and preserving fluorescence. This study aims to compare the efficacy of FOCM to iDISCO in enhancing imaging quality, focusing on vessel intensity and contrast-to-background ratios in one-millimeter-thick mouse brain samples. By comparing one-millimeter-thick mouse brain samples, the study found that FOCM maintained vessel intensity and contrast-tobackground ratios comparable to iDISCO. FOCM was also tested at 5, 60, and 48 hours, demonstrating increased vessel intensity and contrast over time, with 48 hours providing optimal results. This makes FOCM particularly useful for time-sensitive analyses while maintaining image quality

This dissertation explores topics in the economics of higher education. Its goal is to contribute to further understanding of the factors that influence students’ major decisions in college, as well as evaluate the impact of major decisions on the gender wage gap and the types of work available in the college educated labor market.

The first chapter determines the impact of labor market conditions on student majors and, therefore, the composition of the future labor market. The empirical evidence indicates that students choose higher paying majors when they graduate during times of high unemployment. Estimates suggest that students are 2.8% more likely to choose an occupation that pays twice as much in a recession when the unemployment rate is very high but only 2.6% more likely to choose the higher paying major when the unemployment rate is very low. These effects vary by sex, with women being less sensitive to different pay by major at all levels of unemployment.

The second chapter re-examines the extent to which the gender wage gap can be explained by different major decisions made by male and female graduates. Detailed major data leads to the conclusion that inequality in the distribution of majors by sex has increased in the past decade, whereas previously utilized aggregated major data misses this trend. The result is that major can help explain approximately fourteen percent of the gender wage gap, an increase in explanatory power of thirty-five percent compared to less informative data.

The third chapter explores the labor market response to the changing sex composition of potential workers. As women make up an increasing proportion of graduates in a certain field, the average preferences of that labor market may shift for both employers and employees. The result is that, over the time period studied, a ten percent increase in female graduates in a field led to a seven percent decrease in the likelihood that a woman works part-time, and an 8.2% increase in the likelihood that a male graduate does.

Background

Objective

Design

Participants

Approach

Key results

Conclusions

Autonomous robots will soon play a significant role in various domains, such as search-and-rescue, agriculture farms, homes, offices, transportation, and medical surgery, where fast, safe, and optimal response to different situations will be critical. However, to do so, these robots need fast algorithms to plan their motion sequences in real-time with limited perception and battery life. The field of motion planning and control addresses this challenge of coordinating robot motions and enabling them to interact with their environments for performing various challenging tasks under constraints.

Planning algorithms for robot control have a long history ranging from methods with complete to probabilistically complete worst-case theoretical guarantees. However, despite having deep roots in artificial intelligence and robotics, these methods tend to be computationally inefficient in high-dimensional problems. On the other hand, machine learning advancements have led toward systems that can directly perform complex decision-making from raw sensory information. This thesis introduces a new class of planning methods called Neural Motion Planners that emerged from the cross-fertilization of classical motion planning and machine learning techniques. These methods can achieve unprecedented speed and robustness in planning robot motion sequences in complex, cluttered, and partially observable environments. They exhibit worst-case theoretical guarantees and solve a broad range of motion planning problems under geometric collision-avoidance, kinodynamic, non-holonomic, and hard kinematic manifold constraints.

Another challenge towards deploying robots into our natural world is the tedious process of defining objective functions for underlying motion planners and transferring and composing their motion skills into new skills for a combinatorial outburst in robot’s skillset for solving unseen practical problems. To address these challenges, this thesis introduces novel methods, i.e., variational inverse reinforcement learning and compositional reinforcement learning approaches. These methods learn unknown constraint functions and their motion skills directly from expert demonstrations for NMPs and compose them into new complex skills for solving more complicated problems across different domains. Finally, this thesis also presents a model-free neural task planning algorithm that works with never-before-seen objects and generalizes to real world environments. It generates task plans for underlying motion planning and control approaches and solves challenging rearrangement tasks in unknown environments.

Digital financial systems are often positioned by their advocates as imperative to everyday life in Kenya, but inclusion practices embedded in national improvement programs seem to fracture local identities, intentionally and unintentionally creating or exacerbating exclusions even within racial and ethnic groups by defining the conditions of inclusion. This dissertation examines the social consequences of techno-optimistic adoptions of fintech and financial inclusion, which aim to “improve” the international image of the nation through widespread access to formal financial services as part of a National Development Plan. This ambition aims to unite people under financial inclusion and fintech developments. My findings, however, show that entrepreneurs and users of mobile finance solutions feel consistently unsure of the financial inclusion agenda’s expectations, consequently leading many to be financially excluded through various fintech platforms. My study shows how underlying social segmentations—which can be traced through political history and ethnic economic enclaves—are reproduced or exacerbated by the forced adoption of these systems. Financial inclusion may create channels for prosperity and improvement; yet these platforms may engineer debilitating and sometimes precarious financial circumstances for users, borrowers, or entrepreneurs. While financial inclusion and the “fintech revolution” may enable consumers to enjoy greater and more efficient access to financial services, it is important to ensure that inclusion practices do not further financial hardships. Rather, fintechs must generate evidence that their practices are not, in fact, simply rewriting the terms of financial exclusion in their outcomes via alternative means.