Since the recent development of Indium-Gallium-Zinc-Oxide (IGZO) thin-film transistor (TFT), oxide semiconductors have received considerable attention and are widely used as a TFT backplane for the state-of-art active-matrix flat panel display. However, the device application of oxide-TFTs is limited to unipolar devices due to the absence of high-performance p-channel oxide TFTs. This limitation poses a significant challenge in advancing oxide device technology, particularly in realizing complementary metal-oxide semiconductor (CMOS) inverter circuits.Tin monoxide (SnO) emerges as a promising oxide semiconductor to develop high-performance p-channel oxide-TFTs due to relatively good hole mobility and low-temperature processability below 300 °C. But their device performances are still largely behind n-channel oxide TFTs (e.g., IGZO, ZnO) because of high-density subgap defects (1020 cm−3) originated from oxygen vacancy (Vo). By incorporating hydrogen annealing after pulse laser deposition (PLD), Vo induced hole traps in SnO were effectively reduced by forming Sn-H bonds, improving TFT mobility to ~1.8 cm2/Vs and Ion/Ioff ratio to ~105. In addition, the subgap defects were further suppressed by adding back-channel passivation layer, allowing SnO-TFTs to exhibit ambipolar behavior by alleviating the fermi-level pinning near valence band (VB) defects. Lead (Pb2+) exposure is a serious health concern that possesses detrimental effects on public health, causing irreversible damage to minors. Both conventional and home kit detection methods face limitations either high cost or poor reliability and limit of detection (LoD). In addition to studies on oxide semiconductors, I will discuss the label-free graphene field-effect transistor (GFET) based aptamer sensor (aptasensor) specific to Pb2+ sensing. The surface of chemical vapor deposition (CVD) synthesized graphene is functionalized with Pb2+ specific aptamer (Ap), a single-stranded DNA oligonucleotide, serving as a receptor probe binding to linker molecules on graphene. The binding of the target Pb2+ to the Ap is transduced into modulations of the electrical current through graphene FET. The theoretical LoD of GFET aptasensor is discussed based on thermodynamic prediction of probe – linker – channel surface binding energies. The surface density of Ap on graphene is optimized by monitoring minimization of capacitance via electrochemical impedance spectroscopy (EIS) to reach uniform Ap distribution, avoiding steric hinderance due to Ap clustering. The record sensor sensitivity was demonstrated with LoD of ~ 7.0 fM, reaching the LoD predicted from the theoretical thermodynamic prediction.

The objective of this research has been primarily analytical, aiming at a better understanding of why Japan, South Korea, and Taiwan are experiencing different outcomes in their nuclear decisions in the post-Fukushima era and how these deviating outcomes will influence these states’ non-nuclear weapons policies in the coming years. To date, much of the scholarship on nonproliferation in Northeast Asia has paid inadequate attention to the effect of political segmentation and competition within the nuclear policy arena on nuclear decision-making processes and nuclear policies. As Japan, South Korea, and Taiwan democratized and liberalized, the political segmentation within the nuclear policy arena diversified into multiple domestic coalitions with different agendas. Thus, political competition within the nuclear policy arena became more complicated as multiple domestic coalitions interacted and competed for political influence. This research seeks to answer the following questions: What determines the nuclear orientation of Japan, South Korea, and Taiwan in the aftermath of Fukushima? And why are these three states experiencing different outcomes in the post-Fukushima era? Finally, what is the likelihood that these states will reverse their non-nuclear weapons policies in the coming years?

This research argues that in the aftermath of the Fukushima incident, the segmentation and rearrangement of political competition within the nuclear policy arenas of these states are the main factors in determining these states’ nuclear orientations. Nuclear orientation is operationalized via the political behavior of domestic nuclear coalitions which include coalitions that are pro-nuclear energy, pro-nuclear weapons, anti-nuclear energy, and anti-nuclear weapons. Thus, this research contends that in the post-Fukushima era, the final nuclear decision-making of these states is determined by the interplay of these four domestic coalitions within the nuclear policy arena and the ways in which the international and domestic conditions of economy, safety, security, and social norms are filtered through the lenses of these four coalitions. The controlled comparison of these three states in the aftermath of the Fukushima incident provides important benefits for improving our systematic understanding of the relationship between the interplay of coalitions and the nuclear orientation of states.

In the post-Fukushima era, changing international and domestic conditions filtered through the lenses of domestic coalitions affected their nuclear weapons debates differently and resulted in various decision outcomes. These states have been very adamant about their non-nuclear weapons policies while heavily condemning North Korea. As North Korea continued to conduct its nuclear tests, a domino effect or “reactive proliferation,” as many experts predicted, did not occur in Northeast Asia. However, there are still some possibilities that reactive proliferation could occur in Northeast Asia and spill over to other regions if any one of these states decides to go nuclear in the future.

Taiwan is moving toward complete denuclearization by removing its civilian nuclear programs. South Korea is gradually moving toward complete denuclearization via a gradual phase out of its nuclear power and finding different paths to complete its nuclear fuel cycle. However, Japan’s nuclear orientation is circling back to the original position that it had prior to the Fukushima incident. According to the findings of the case study chapters, this dissertation cautiously envisions that, for different domestic political reasons, Japan and South Korea are more prone to go nuclear than Taiwan if the U.S. nuclear umbrella fails to work properly in the coming years. Japan has an ambition to become the power house of Asia once more. Thus, nuclear weapons might not be an end goal but a necessary step on its way to becoming a great power. This study contends that Japan is more prone to go nuclear due to its political motivations and the consistency shown by its leadership on the matter of nuclear hedging throughout the years. Unlike South Korea and Taiwan, the Japanese leadership continuously used external threats, such as China and North Korea, to rouse nationalistic sentiment within the general population and to justify its remilitarization process. In particular, the surge of nationalism in Japan should be carefully monitored because this will not influence its short-term, but will influence its long-term national strategy. In contrast, this study contends that South Korea is prone to go nuclear due to high public support for nuclear weapons. Even though public support for nuclear weapons is showing a pattern of downward trend since 1999, the idea of acquiring nuclear weapons is still popular among many South Koreans. The recent polls from 2017 to 2018 vary from 43% to 67%. Thus, even though there are no immediate concerns for these states to abandon their non-nuclear weapons policies, the international community needs to keep close eyes on the public support for the nuclear weapons in South Korea and the surge of nationalism and the remilitarization process that is currently in progress in Japan.

For a robot to interact with its environment, it must perceive the world and understand how the world evolves as a consequence of its actions. This thesis studies a few methods that a robot can use to respond to its observations, with a focus on instances that can leverage visual dynamic models. In general, these are models of how the visual observations of a robot evolves as a consequence of its actions. This could be in the form of predictive models that directly predict the future in the space of image pixels, in the space of visual features extracted from these images, or in the space of compact learned latent representations. The three instances that this thesis studies are in the context of visual servoing, visual planning, and representation learning for reinforcement learning. In the first case, we combine learned visual features with learning single-step predictive dynamics models and reinforcement learning to learn visual servoing mechanisms. In the second case, we use a deterministic multi-step video prediction model to achieve various manipulation tasks through visual planning. In addition, we show that conventional video prediction models are unequipped to model uncertainty and multiple futures, which could limit the planning capabilities of the robot. To address this, we propose a stochastic video prediction model that is trained with a combination of variational losses, adversarial losses, and perceptual losses, and show that this model can predict futures that are more realistic, diverse, and accurate. Unlike the first two cases, in which the dynamics model is used to make predictions for decision-making, the third case learns the model solely for representation learning. We learn a stochastic sequential latent variable model to learn a latent representation, and then use it as an intermediate representation for reinforcement learning. We show that this approach improves final performance and sample efficiency.