DSpace@RPI (Rensselaer Polytechnic Institute)
Not a member yet
6809 research outputs found
Sort by
Semantic Technologies for Clinically Relevant Personal Health Applications
No full textDespite recent advances in digital health solutions and machine learning, personal health applications that aim to modify health behaviors are still limited in their ability to offer more personalized decision support. Moreover, while many personal health applications cater to general health and well-being, there remains a significant opportunity to increase the clinical relevance of the insights being generated. This chapter describes the motivation for, and illustrative applications of, semantic technologies for enabling clinically relevant personal health applications. We present two use cases that demonstrate how semantic web technologies, in combination with machine learning and data mining methods, can be used to provide personalized insights to support behaviors that are consistent with nutritional guidelines for people with diabetes
Development of in-situ FIB/SEM techniques for microstructural characterization and evolution of polycrystalline metals
No full textDecember 2021School of EngineeringA fundamental understanding of materials processing through thermal and mechanical means is a significant factor when choosing materials to use across industries. The ability to control the failure of materials, or to choose a material strength that is appropriate for different applications, is useful across many industries. Current methods of testing for desired material properties include multiple time-consuming steps. For example – in order to inspect the strength of a metal, one may have to go through multiple heat treatments of the metal, inspection, sample preparation, mechanical testing, and final failure analysis. A faster and more reliable approach that could be universally used to help reduce the time of this testing would be valuable from an industry standpoint. In order to assist in the reduction of materials processing time, a more thorough understanding of the development of microstructure within material systems can be developed by observing microstructure evolution in real time under different thermal and mechanical processes. In this work, we achieve this in two ways: 1) The investigation of copper grain growth under uniform and non-uniform temperature gradients and 2) The study of Ti 6%Al 4%V (Ti 6-4) alloys during thermomechanical processing and cooling. To achieve this, the development of novel techniques, using a scanning electron microscope integrated with both thermal (Cu) and thermomechanical (Ti 6-4) processing techniques are created for each material system. To investigate grain growth in copper, an in-situ heater is employed to allow samples to reach temperatures from 300-500 °C for observation. This is then coupled with a micro-heater array allowing the observation of copper thin films deposited over ten heater array lines, capable of creating a temperature gradient across a sample. This novel technique allows up to a ~30 °C/cm temperature gradient across the sample. This allows in principle for investigation, not only of thin film copper grain growth at uniform temperatures over time, but for the modification of the gradients to optimize the grain structure. Ti 6%Al 4%V is a two-phase alloy commonly used in aerospace applications due to its ability to maintain high strength at service temperatures in excess of 500 °C. The resulting microstructures of Ti 6%Al 4%V span equiaxed, fully lamellar, martensitic and bi-modal microstructures. Each can be obtained through a combination of multiple deformation, recrystallization and annealing steps to obtain the final microstructure desired. The ability to deconvolute the effects of different processing steps (homogenization, deformation, recrystallization, and annealing) and combine them in new ways could reduce processing time, if the optimal microstructure can be achieved with fewer thermomechanical treatments. A stage that is capable of thermomechanical processing via localized heating and tensile testing for use inside of a scanning electron microscope was commissioned and modified, with capabilities of heating the sample above the Ti 6%Al 4%V β transus temperature of 995 °C, and straining the sample in ranges from 1x10-3 s-1 to 1x10-5 s-1. In addition, the capability of monitoring cooling rates in real time is developed. We aim to maintain uniform fields of view and imaging conditions in order to assess microstructure as thermomechanical deformation occurs. A series of experiments is developed to deconvolute the impact of cooling rate on lamellar microstructure formations. This leads to understanding the mechanisms of lamellar microstructure formation, and the optimization of the size of lamellae for future materials processing. Integrated indentation capabilities being developed will also allow for the potential of real time feedback into the microstructural properties of the lamellar structure, in order to further inform the optimal size of the microstructure. The resulting microstructure library leads to the ability to correlate multiple parameters with microstructure at different processing steps, resulting in a clearer understanding of the influence of those steps. This has the potential to contribute to the optimization of materials, significantly reduce the time needed to process Ti 6%Al 4%V and other alloys, and give insight on the fundamental growth of microstructure.DEn
Optimization of controlled environment agriculture (CEA) for growing crops
No full textMay 2022School of EngineeringControlled environment agriculture (CEA) is a rapidly growing production system that promotes food security, environmental stewardship, and efficient resource use. According to a recent report, global market for CEA is projected to grow at an 18.7 % annual rate and reach $172 billion in 2025. In this work, the agricultural output (crop yield and nutrient content) of three kale cultivars (‘Toscano’, ‘Redbor’, and ‘Winterbor’) grown in environmental chambers was compared with the output of identical varieties grown in the field and greenhouse systems. Differences in morphology and growth kinetics were observed between the different systems. Overall, higher phytochemical content was observed in plants from the growth chamber environment. At seedling stage, lutein concentration (macular pigment) in leaves from growth chamber was 37 – 72 % higher than lutein from other growth systems. In this study, cultivar type and developmental stage were also found to be important factors that determine nutritional quality in kale. Composition of light spectra is known to influence the morphology of plants and leaf phytochemicals. In the second study, the effect of different blue peak wavelength (400 – 450 nm) on kale development was investigated using tunable LED lamps inside CEA units. Pigment content was observed to be influenced by the type and irradiance level of blue spectrum during growth. Linear increase in lutein and chlorophyll a concentration was found as a function of blue peak wavelength. Finally, in the third study, a cost-effective spectrophotometric method for the quantification of pigments in leaf extracts was developed. Harnessing the spectral and physico-chemical properties of the principal light-absorbing compounds (chlorophylls and carotenoids) found in leafy greens, simultaneous equations were formulated and tested. Overall, the research findings will contribute to the adoption of CEA systems for growing nutritious crops.Ph
Next-generation anode materials for metal-ion batteries
No full textAugust 2021School of EngineeringLithium-ion battery technology is one of the greener alternatives of non-renewable energy sources for overcoming burgeoning energy demand. However, current-state-of-art lithium-ion batteries (LIBs) face critical challenges among which low performance, scarcity, and uneven distribution of lithium metal in the earth’s crust are major ones. The low-performance challenge can be overcome by utilizing alloy-based or conversion-based chemistries instead of intercalation/de-intercalation chemistry, used in a current-state-of-art LIBs, to store a large number of lithium atoms in one cycle. Scarcity challenges can be overcome by exploring beyond LIBs and two important candidates are sodium-ion batteries (SIBs) and potassium-ion batteries (KIBs). Therefore, various alloy-based and conversion-based materials have been explored not only for the fabrication of efficient lithium-ion but also for sodium-ion and potassium-ion batteries. Constant development in the fabrication of innovative materials led to the introduction of various new innovative materials which has not been explored before for battery application. Tellurene and phosphorene are two such materials. In this Ph.D. work, I have explored tellurene and phosphorene as electrode materials for the fabrication of all three alkali-ion batteries. In the first aspect of my work, we fabricated tellurene (Te) based LIBs, SIBs, and KIBs. Through density functional theory calculations (DFT) and in-situ transmission electron microscopy (TEM) technique, we confirmed that Te when reacts with lithium (Li), sodium (Na), and potassium (K) led to the formation of A2Te type product where A is either Li or Na or K. Significant difference in crystal property of product was observed during in-situ TEM measurement wherein Li2Te showed single-crystal property whereas Na2Te and K2Te showed polycrystalline nature. Nudged elastic band calculations (NEB) and ab-initio molecular dynamics (MD) confirmed isotropic diffusion of Li atoms on the surface and inside few-layer Te sheets. However, the diffusion of Na atoms and K atoms was highly anisotropic. This allowed us to present a hypothesis stating that if the diffusion of alkali atoms is isotropic, the reaction between Te and alkali atoms will happen simultaneously in all directions due to which no grain boundaries will form during the reaction. Te-based LIBs were able to deliver a moderate initial capacity of ~ 350 mAh g-1, higher than Te-based SIBs ( ~ 160 mAh g-1) and Te-based KIBs (~ 150 mAh g-1). However, huge volume expansion in Te during chemical reaction led to pulverization/delamination of Te electrode due to which capacity degradation was observed in all three alkali-ion batteries. Te-based LIBs were able to deliver a stable capacity of ~ 120 mAh g-1 after 50 cycles proving that gravimetrically Te does not provide any advantage over conventional graphite used in current-state-of-art LIBs. However, the volumetric picture was different wherein Te-based LIBs were able to deliver a very high initial capacity of ~ 1800 mAh cm-3 with a stable capacity of ~ 700 mAh cm-3 (2x graphite) after 50 cycles. Pulverization/delamination was still observed and to overcome this issue, graphene coating over Te was used. Te was able to retain over ~ 85% of its initial capacity after 100 cycles in presence of graphene. In the second aspect of my work, I have explored phosphorene for battery application. Before us, phosphorene has already been explored for LIBs and SIBs and delivers a very high capacity of ~ 2600 mAh g-1 corresponding to Li3P/Na3P alloy formation. Hence, we explore phosphorene for only KIBs with an expectation that it will lead to a similar type of alloy product. Phosphorene and phosphorene carbon composite was synthesis through a liquid-phase exfoliation method. Two carbon-based materials were utilized, reduced graphene oxide (rGO) and single-walled carbon nanotubes (sCNTs), to overcome the pulverization/delamination challenge. Phosphorene was able to deliver a very high initial capacity of ~ 1200 mAh g-1 but rapid capacity decay to ~ 50 mAh g-1 was observed in subsequent cycles. rGO and sCNTs were used to provide mechanical buttress to the electrode wherein rGO proved to be a better candidate in preventing delamination of phosphorene electrode because of its similar dimensionality with phosphorene. Few-layer phosphorene/reduced graphene oxide (FLP-rGO) showed a stable capacity of ~ 550 mAh g-1 after 50 cycles with ~ 97% columbic efficiency. Using DFT calculations, ex-situ X-ray photoelectron spectroscopy, and X-ray diffraction techniques, we confirmed that phosphorene when reacts with K led to the formation of K4P3 alloy. FLP-rGO performed well even in a full-cell configuration while utilizing spherical potassium cobalt oxide (s-KCO) as a cathode. Throughout my Ph.D. work, we have utilized nano-sized tellurene and phosphorene for the fabrication of alkali-ion batteries. The long-standing debate over utilizing nano-sized over micro-sized electrodes for battery fabrication concludes that battery engineers are reluctant in utilizing nano-sized material due to its low first cyclic columbic efficiency, low volumetric capacity, high fabrication cost, and complex experimental synthesis. Alloy or conversion-based materials were expected to work better in their nano-sized form, however, a growing body of works indicates that micro-sized material can be utilized with proper electrode design, binder modification, and electrolyte engineering. Therefore, micro-sized tellurene and phosphorene can potentially be explored in the battery field to enhance their industrial significance.Ph
Guiding principles for technical infrastructure to support computable biomedical knowledge
No full textOver the past 4 years, the authors have participated as members of the Mobilizing Computable Biomedical Knowledge Technical Infrastructure working group and focused on conceptualizing the infrastructure required to use computable biomedical knowledge. Here, we summarize our thoughts and lay the foundation for future work in the development of CBK infrastructure, including: explaining the difference between computable knowledge and data, and contextualizing the conversation with the Learning Health Systems and the FAIR principles. Specifically, we provide three guiding principles to advance the development of CBK infrastructure: (a) Promote interoperable systems for data and knowledge to be findable, accessible, interoperable, and reusable. (b) Enable stable, trustworthy knowledge representations that are human and machine readable. (c) Computable knowledge resources should, when possible, be open. Standards supporting computable knowledge infrastructures must be open.American Association for Cancer Research; International Business Machines Corporation; National Association of Attorneys General; National Cancer Institute; National Center for Advancing Translational Sciences; National Institute of Environmental Health Sciences; National Institute on Drug Abuse; National U.S. Science Foundation; Patient-Centered Outcomes Research Institute; Truth Initiative; U.S. National Library of Medicin
A Theoretically Grounded Benchmark for Evaluating Machine Commonsense
No full textProgramming machines with commonsense reasoning (CSR) abilities is a longstanding challenge in the Artificial Intelligence community. Current CSR benchmarks use multiple-choice (and in relatively fewer cases, generative) question-answering instances to evaluate machine commonsense. Recent progress in transformer-based language representation models suggest that considerable progress has been made on existing benchmarks. However, although tens of CSR benchmarks currently exist, and are growing, it is not evident that the full suite of commonsense capabilities have been systematically evaluated. Furthermore, there are doubts about whether language models are 'fitting' to a benchmark dataset's training partition by picking up on subtle, but normatively irrelevant (at least for CSR), statistical features to achieve good performance on the testing partition. To address these challenges, we propose a benchmark called Theoretically-Grounded Commonsense Reasoning (TG-CSR) that is also based on discriminative question answering, but with questions designed to evaluate diverse aspects of commonsense, such as space, time, and world states. TG-CSR is based on a subset of commonsense categories first proposed as a viable theory of commonsense by Gordon and Hobbs. The benchmark is also designed to be few-shot (and in the future, zero-shot), with only a few training and validation examples provided. This report discusses the structure and construction of the benchmark. Preliminary results suggest that the benchmark is challenging even for advanced language representation models designed for discriminative CSR question answering tasks.
Benchmark access and leaderboard: this https URL Benchmark website: this https UR
Low-order methods for nonconvex functional constrained optimization
No full textJuly 2022School of ScienceRecently, many real-world problems in engineering and data science not only have very large scales and complicated functional constraints, but also go beyond the scope of convex optimization and inevitably include nonconvex structures. This thesis focuses on developing and analyzing low-order methods for nonconvex functional constrained optimization. In this thesis, I propose several low-order methods, and analyze the complexity of the proposed methods for finding near-KKT points of nonconvex composite problems with either convex or nonconvex functional constraints. All proposed methods generally combine the frameworks of the augmented Lagrangian method, the proximal point method, and my designed subroutines to solve certain unconstrained subproblems. The best-known complexity results are established to all proposed methods on corresponding classes of problems. Numerical experiments demonstrate the efficiency of the proposed methods on a large number of both classical optimization problems and real-world machine learning examples.Ph
Optical navigation using quadrics and indirect indexing & search methods
No full textMay 2022School of EngineeringAutonomy plays a crucial role in navigation to other celestial bodies and around our own planet. With steadily increasing demand for radiometric data on an already strained network, autonomous decision-making is becoming evermore central to spacecraft navigation. Autonomy relies on external sensory data, including digital imagery. While images certainly contain invaluable information, they possess inherent limitations---including scale ambiguities and perspective distortions. Leveraging projective geometry and invariant theory---both vital foundations to computer vision---this work resolves particular questions in navigation by exploiting geometric properties using invariant theory, namely: boundary localization, invariant data storage and interrogation, and pose estimation.Ph
A methodology to evaluate equity in resource accessibility
No full textAugust 2022School of EngineeringInequalities present in society have been of increased interest in recent years, particularly in the United States. One source of inequality is unequal access to important resources, such as shopping. Recent developments in shopping mean that the ways individuals obtain goods have changed significantly in the past decade, with little research into some of these new developments. Many of these developments concern online shopping, which has risen from a minor factor to a large portion of the retail economy in the past two decades. The COVID-19 pandemic forced people to change how they shop, highlighted existing inequities in shopping access, and resulted in permanent changes to how families across the world obtain goods. Prior studies have indicated large discrepancies in accessibility to shopping, with focus on the relationship between economic status and accessibility. Other studies have touched on the lower accessibility to shopping in rural areas, particularly concerning the area of home delivery. However, large gaps in the literature exist, with online shopping being relatively understudied and the unique challenges to rural areas being understudied. This dissertation presents a methodology to access accessibility to online and in person shopping. Using a survey conducted by the author, econometric models are generated to determine the factors influencing accessibility to five types of online shopping and four types of in person shopping. These models are then applied to a case study of eastern Upstate New York, determining the accessibility to each type of shopping in every ZIP Code in the study area and generating composite accessibility scores. These accessibility scores indicate a wide variation in both in person and online shopping accessibility across the study area, with urban and high-income areas more likely to have high accessibility to shopping than low-income and rural areas. Policies to remedy the inequities evidenced by the case study are then presented, with discussion of how they may be implemented.Ph
Acoustical analysis of coupled spaces based on spatial parameters
No full textAugust 2022School of ArchitectureIn the field of Architectural Acoustics, there has been a significant focus on binaural hearingand the way room design impacts a listener’s experience within that space. Up to this point,
the Interaural Cross-correlation Coefficient (IACC) and its related property, Interaural Decorrelation
Coefficient (IADC), have been integral in the understanding of how the sound
envelops a listener. As previous research has indicated, it is possible to analyze single volume
spaces using the IADC method with a 90 ms dividing value for the late reverberation tail.
The aim of this project is to study the IADC within the context of a coupled volume situation
in order to answer the following questions. Where does the threshold between early and late
sound energy occur in a coupled situation? Is it possible to simulate the experience of
being in a coupled space by manipulating the late reverberation tail? Is there a perceivable
difference in the acoustic performance of coupled spaces when an aperture is closed versus
when it is open? In order to answer these questions, there must be an experiment within a
coupled space, as well as some level of perceptual testing. This thesis uses a binaural dummy
head within the main volume for the examination of the sound field around a human subject
when exposed to the secondary volume via an aperture. The research in this thesis indicates
that the IADC in the later reverberation tails may vary drastically within coupled volume
acoustics. Additional study may be required to find out how to best distinguish between early
and late energy in a coupled space. Future research may support the continued construction
of coupled volume concert venues as a way to increase the acoustical potential of these spaces.M