DSpace@RPI
DSpace@RPI is a repository of Rensselaer Polytechnic Institute's theses and dissertations which are available in digital format, largely from 2006 to present, along with other selected resources.
Recent Submissions
Item A Community-driven vision for a new Knowledge Resource for AI(arXiv, 2025-06-19)The long-standing goal of creating a comprehensive, multi-purpose knowledge resource, reminiscent of the 1984 Cyc project, still persists in AI. Despite the success of knowledge resources like WordNet, ConceptNet, Wolfram|Alpha and other commercial knowledge graphs, verifiable, general-purpose widely available sources of knowledge remain a critical deficiency in AI infrastructure. Large language models struggle due to knowledge gaps; robotic planning lacks necessary world knowledge; and the detection of factually false information relies heavily on human expertise. What kind of knowledge resource is most needed in AI today? How can modern technology shape its development and evaluation? A recent AAAI workshop gathered over 50 researchers to explore these questions. This paper synthesizes our findings and outlines a community-driven vision for a new knowledge infrastructure. In addition to leveraging contemporary advances in knowledge representation and reasoning, one promising idea is to build an open engineering framework to exploit knowledge modules effectively within the context of practical applications. Such a framework should include sets of conventions and social structures that are adopted by contributors.Item Commonsense AI in the History of the Web(Association for Computing Machinery, 2025-05-08)Machine common sense (MCS)-the challenge of enabling computers to grasp everyday human knowledge-has been a grand challenge in Artificial Intelligence (AI) since the 1950s. While recent advances in large language models have led to impressive progress, there is still no consensus on how much common sense today's AI actually possesses. In this brief review, we revisit the historical development of MCS in the context of the Web, examining how the Web's evolution-from early knowledge representation efforts to knowledge graphs, the Semantic Web, and crowdsourcing-has shaped MCS research. We argue that key breakthroughs in Web technologies were instrumental in addressing longstanding challenges of scale and coverage in commonsense reasoning. At the same time, MCS research has influenced the development of core Web applications, including intelligent agents, plausibility-based reasoning, and robust evaluation of black-box AI systems.Item Synthesis and characterization of monomeric ruthenium-based catalysts for water oxidation(Rensselaer Polytechnic Institute, Troy, NY, 2025-05)Nature converts solar energy into chemical energy (in the form of carbohydrates), releases dioxygen, and fixes carbon dioxide through photosynthesis. Light-driven water oxidation, one of the most energetically demanding reactions in nature (2H2O → O2 + 4H+ + 4e-), occurs in the photosynthetic reaction center, photosystem II (PSII). The structure of PSII revealed the tetranuclear manganese-calcium-oxo (Mn4Ca-oxo) cluster in the oxygen-evolving complex (OEC) that is known to catalytically initiate the water oxidation to produce protons, electrons and the release of dioxygen. Inspired by the catalytic manganese-calcium-oxo (Mn4Ca-oxo) cluster, several catalysts for artificial water oxidation have been developed with varied metal centers. The scope of the metals includes the first-row transition metals, ruthenium, and iridium. In Chapter 1, we describe the water oxidation reaction in Nature and progress on the development of artificial water oxidation complexes over the past few decades. Among the aforementioned metals, ruthenium-based catalysts have been studied extensively to explore the structure-activity relationships in the hope of illuminating a strategy for designing an efficient catalyst for water oxidation. Based on the current molecular ruthenium models for water oxidation, we designed, synthesized and characterized a series of ruthenium complexes with a negatively charged dicarboxylate backbone. Chapter 2 describes the synthesis and characterization of ruthenium-based complexes with symmetric backbone ligands, such as, pda2− (2,6-pyridinyldiacetate), pba2− (pyridine-2,6-bis(α-oxo) acetate) and pdc2− (2,6-pyridinedicarboxylate) with various ancillary ligands, tfmp, py, pic and dmap, differing in electron-donating ability. 1H NMR was employed to investigate the electronic effect of ancillary ligands on the protons from the backbone ligands and those from the coordinated ancillary ligand. UV-Vis spectroscopy was used to study the electronic absorption of the Ru-pba family. The UV-Vis spectra revealed that the electron-donating ability of ancillary ligands can affect the metal-ligand charge transfer (MLCT) bands. The 1H NMR spectra revealed that a stronger electron-donating ancillary ligand will result in an upfield shift of the peaks. Chapter 3 describes the synthesis and characterization of ruthenium-based complexes with asymmetric backbone ligands, such as, cmpc2− (6-(carboxymethyl)-pyridine-2-carboxylate) and cpa2− (6-carboxy-α-oxo-2-pyridine acetate) with various ancillary ligands, tfmp, py, pic and dmap, differing in electron-donating ability. 1H NMR spectroscopy was employed to investigate the electronic effect of the ancillary ligands on the protons from the backbone ligands and those from the coordinated ancillary ligand. UV-Vis spectroscopy was used to study the electronic absorption of the Ru-cpa family. Comparison of the 1H NMR and UV-Vis spectra are presented in this chapter. The crystal structure of the complexes revealed an O-Ru-O bite angle ranging from 172° – 178° in the complexes with little distortion of the octahedral configuration, indicating increased stability and ease of access to bind water molecules at the metal center. Functional catalytic studies of these complexes in the future will contribute insight on the structure-activity relationships and serve as a motivation for the design of novel molecular catalysts for water oxidation.Item Deep probabilistic and generative models for x-ray based imaging and ecg(Rensselaer Polytechnic Institute, Troy, NY, 2025-05)Artificial intelligence (AI) is poised to transform modern medicine and, in particular, medical imaging, as the need for healthcare services continues to outstrip available resources worldwide. Deep learning models show promise in enhancing clinical workflows and patient outcomes through data-driven diagnoses and improved image quality. Despite the pressing need and considerable research efforts, several challenges have delayed the integration of AI into healthcare: Neural networks often lack explainability, making overconfident inferences on out-of-sample data, and are susceptible to adversarial attacks—small, targeted input perturbations capable of fooling otherwise accurate networks. Furthermore, large models are data-hungry, yet patient images and information are legally guarded by health privacy protections. Future virtual clinical trials for medical device validation also require high-quality synthetic datasets that sufficiently represent rare pathologies and population demographics. While deep generative models may address these data gaps, their outputs often lack clinical fidelity despite appearing visually convincing. Finally, many image-enhancement tasks, such as deblurring, lack the ground-truth labels necessary for supervised learning, forcing reliance on simulated degradations that can introduce artifacts when applied to real-world data. Compounding the above issues is the high-dimensional nature of most medical signals, invalidating solutions that cannot be feasibly scaled. The following dissertation investigates solutions to several of the aforementioned challenges within specific applications, primarily leveraging deep probabilistic and generative models. First, we examine how diverse deep ensembles, aided by a feature decorrelation mechanism, can improve adversarial robustness in high-dimensional tasks like electrocardiogram classification. Next, drawing inspired by the 2023 AAPM Deep Generative Modeling Challenge, we employ denoising diffusion probabilistic models to produce synthetic medical images that are realistic in both visual appearance and clinical relevance—an effort that earned first place in the competition. Finally, we adapt state-of-the-art simulation techniques to create realistic, system specific degradations to train deep deblurring models for photon-counting computed tomography. By scaling a diffusion model to 3D through a joint 2D inference process and disentangling noise from signal prior to deblurring, we successfully mitigate texture distortions and improve performance on real-world data.Item Exploring efficacy in digital therapeutics: serious games for theory of mind training and visual rehabilitation(Rensselaer Polytechnic Institute, Troy, NY, 2025-05)As digital media increasingly transforms healthcare and education, the unique affordances of serious video games present distinct challenges and opportunities, setting them apart from traditional serious games. While conventional serious games often superficially gamify clinical practices, diminishing meaningful engagement, or replicate clinical protocols so rigidly that intrinsic player motivation suffers, serious video games uniquely offer interactive affordances that could effectively reconcile clinical precision with authentic player engagement. This dissertation examines this critical tension through the iterative development and empirical evaluation of two digital therapeutic prototypes: \textit{Emotion Adventure}, designed to foster Theory of Mind, the ability to understand and interpret others’ emotional and mental states, in children with Autism Spectrum Disorder, and \textit{Eye Rehab}, a virtual reality game aimed at improving stroke-related visual impairments. These prototypes were systematically designed and evaluated using the Mechanics, Dynamics, and Aesthetics (MDA) framework, which methodically connects foundational game mechanics, emergent player dynamics, and experiential aesthetics to ensure balanced game design. In usability evaluations, \textit{Emotion Adventure} employed a narrative-driven approach to successfully promote empathic decision-making and maintain player engagement within structured gameplay interactions. However, given the complexities and resource demands involved in empirically measuring cognitive therapeutic outcomes, a second prototype, \textit{Eye Rehab}, was developed. This physiological digital therapeutic utilized virtual reality-based gaze interactions to precisely target measurable improvements in visual alignment and ocular motor functions, validated clinically through the Lancaster Red-Green test, a standardized diagnostic tool used to assess ocular alignment and muscle function. Building upon insights gained through these prototypes, this dissertation hypothesizes a replicable design approach termed \textit{Selective Simulation}, which strategically embeds essential therapeutic actions directly into core gameplay mechanics. Unlike earlier theoretical concepts such as persuasive or applied games that offer generalized guidance, \textit{Selective Simulation} provides concrete and empirically informed design principles to intentionally integrate therapeutic activities within engaging game mechanics. Ultimately, this dissertation contributes to the broader field by proposing a replicable and structured framework for serious video game design, bridging theoretical insights from media studies, cognitive psychology, and human-computer interaction with methodologies rooted in clinical practice. This interdisciplinary approach underscores the distinct potential and complexity of video games as digital therapeutics, advocating for designs that rigorously balance therapeutic efficacy and engaging gameplay.
Communities in DSpace@RPI
Select a community to browse its collections.