Recent publications
This study investigates the development and evaluation of a Retrieval-Augmented Generation (RAG)-based statistics tutor designed to assist students with quantitative analysis methods. The RAG approach was employed to address the well-documented issue of hallucination in Large Language Models (LLMs). A computer tutor was developed that utilizes ChatGPT to understand student questions in natural language and retrieves relevant information from validated course notes to generate responses. The tutor was evaluated using twenty questions curated from the course notes. Two instructors who taught the course using the course notes assessed the quality of the tutor’s response using a pre-defined rubric. The evaluation results indicate that the RAG-based tutor was able to provide correct and relevant answers to all twenty questions and to generate R code snippets upon request. This study highlights the potential of RAG-based tutors to provide accurate and personalized learning experiences for students while mitigating the risk of providing fabricated information often associated with LLMs.
By serving as an analog to traffic signal lights, communication signaling for drone to drone communications holds the key to the success of advanced air mobility (AAM) in both urban and rural settings. Deployment of AAM applications such as air taxis and air ambulances, especially at large-scale, requires a reliable channel for a point-to-point and broadcast communication between two or more aircraft. Achieving such high reliability, in a highly mobile environment, requires communication systems designed for agility and efficiency. This paper presents the foundations for establishing and maintaining a reliable communication channel among multiple aircraft in unique AAM settings. Subsequently, it presents concepts and results on wireless coverage and mobility for AAM services using cellular networks as a ground network infrastructure. Finally, we analyze the wireless localization performance at 3D AAM corridors when cellular networks are utilized, considering different corridor heights and base station densities. We highlight future research directions and open problems to improve wireless coverage and localization throughout the manuscript
In this work, we performed a detailed analysis of the x-ray photoemission spectroscopy (XPS) of the Mn 2p peak for Mn3O4(001) thin films. This is a challenging task since Mn3O4 is composed of two different cations, Mn²⁺ at tetrahedral and Mn³⁺ at octahedral sites, which both contribute to the XPS spectra. The oxide spectra consist of many multiplets arising from the angular momentum coupling of the open Mn 2p and 3d shells, thus increasing the spectrums’ complexity. Moreover, the energy spacing and intensities of the different multiplets also reflect the covalent mixing between Mn 3d and O 2p shells. However, we show that a detailed analysis, which provides relevant information about the cations in the oxide structure, is possible. We prepared experimentally different Mn3O4 films on Au(111), and their structure was monitored with the diffraction pattern obtained with low-energy electron diffraction. The Mn 2p spectra were fit, guided by cluster model theoretical predictions, and checked for films prepared at different oxygen partial pressures. Therefore, we could observe the Mn²⁺ and Mn³⁺ cations’ relative concentration in the Mn 2p mains peaks.
We conducted a multilevel meta-analysis of 390 effect sizes from 167 studies with 157,923 participants examining the relationship between connectedness with lesbian, gay, bisexual, transgender, and queer/questioning (LGBTQ+) communities and health-related outcomes, following Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. We conducted our initial search in January 2023 in APA PsycInfo, ERIC, Medline, and Open Dissertations, selecting studies that (a) measured LGBTQ+ community connectedness, (b) measured health, and (c) provided an estimate of the relationship between LGBTQ+ community connectedness and health. We found that connectedness with LGBTQ+ communities promotes mental health (r = .11), well-being (r = .17), and physical health (r = .09). Conversely, we found that connectedness with LGBTQ+ communities promotes substance use among younger participants, likely through behavioral engagement with LGBTQ+ others. We found that connectedness with LGBTQ+ communities was related to less mental health and more suicidality for younger people, likely because younger LGBTQ+ people seek out connectedness in response to this psychological distress. We also found that connectedness was not as health-promoting for LGBTQ+ individuals with multiple marginalized identities and that psychological feelings of belongingness with LGBTQ+ communities are generally more health-promoting than behavioral community engagement. Results from a narrative review and moderation meta-analyses suggested that, contrary to predictions made by minority stress theory, connectedness with LGBTQ+ communities does not buffer the relationship between minority stressors and health. Rather, meta-analytic mediation analyses suggested that proximal minority stressors negatively impact health-related outcomes by reducing connectedness with LGBTQ+ communities and that distal minority stressors are often less impactful on health-related outcomes because they promote connectedness with LGBTQ+ communities.
Unlike naturally occurring oxide crystals such as ruby and gemstones, there are no naturally occurring nitride crystals because the triple bond of the nitrogen molecule is one of the strongest bonds in nature. Here, we report that when the transition metal scandium is subjected to molecular nitrogen, it self-catalyzes to break the nitrogen triple bond to form highly crystalline layers of ScN, a semiconductor. This reaction proceeds even at room temperature. Self-activated ScN films have a twin cubic crystal structure, atomic layering, and electronic and optical properties comparable to plasma-based methods. We extend our research to showcase Sc’s scavenging effect and demonstrate self-activated ScN growth under various growth conditions and on technologically significant substrates, such as 6H–SiC, AlN, and GaN. Ab initio calculations elucidate an energetically efficient pathway for the self-activated growth of crystalline ScN films from molecular N2. The findings open a new pathway to ultralow-energy synthesis of crystalline nitride semiconductor layers and beyond.
Previous literature has well‐established genetic factors as being associated with neuroblastoma (NB). About 1%–2% of NB cases are familial, with 85% of these cases predisposed to mutations in the PHOX2B and ALK genes. The genetic basis of sporadic NB has been studied through genome‐wide association studies and next‐generation sequencing approaches. Particularly, germline variants, as well as copy number variations, confer increased risks of NB, often with effect estimates ≥1.5, underscoring the strong genetic contributions to NB. However, the strength of the association varied in non‐genetic factors. Some risk factors, such as birth defects, maternal illicit drug use, and early infections, had relatively stronger associations (effect estimates ≥1.5 or ≤0.67), while some other factors remain inconclusive. This suggests that certain non‐genetic factors may play a more prominent role in NB risk, while further research is needed to clarify the impact of others. We synthesized and critically evaluated existing literature on the risk factors of NB to provide an overview, analyze the current state of knowledge, and outline a research path to address the relative contributions of genetic and non‐genetic factors in NB. Future epidemiologic studies should incorporate novel methods for measuring genetic and non‐genetic factors to comprehensively assess the full extent of factors contributing to NB. Furthermore, the utilization of dried blood spots holds promise to overcome technical and recruitment challenges for future studies. These strategies will contribute to a more holistic understanding of NB etiology and potentially lead to improved prevention strategies.
α‐substituted ketones are important chemical targets as synthetic intermediates as well as functionalities in in natural products and pharmaceuticals. We report the α‐acetylation of C(sp3)‐H substrates R‐H with arylmethyl ketones ArC(O)Me to provide α‐alkylated ketones ArC(O)CH2R at RT with tBuOOtBu as oxidant via copper(I) β‐diketiminato catalysts. Proceeding via alkyl radicals R•, this method enables α‐substitution with bulky substituents without competing elimination that occurs in more traditional alkylation reactions between enolates and alkyl electrophiles. DFT studies suggest the intermediacy of copper(II) enolates [CuII](CH2C(O)Ar) that capture alkyl radicals R• to give R‐CH2C(O)Ar under competing dimerization of the copper(II) enolate to give the 1,4‐diketone ArC(O)CH2CH2C(O)Ar.
The field of psychological injury and law is marked by use of psychometrically sound validity tests that use empirically derived cut scores to determine the credibility of cognitive deficits and psychological symptoms in forensic and related disability assessments (FDRA). Performance validity tests (PVTs) are used in neuropsychological/cognitive assessments to determine the extent to which test scores reflect true ability levels. Symptom validity tests (SVTs) are designed to evaluate the credibility of self-reported level of excessive report in behaviors, emotions, and thoughts. They monitor the rate of endorsement of rare, absurd, impossible, and improbable symptoms. The authors argue for a 30% rule as a tentative multivariate threshold for invalid presentation (with provisos). In other words, failure on about one third of the PVTs/SVTs administered should be required before deeming the overall profile non-credible to control for the threat of inflated false positive error due to the increasing number of instruments used. Typically, workers in the field use the multivariate threshold of ≥ 2 PVT failures in FDRA to deem an entire profile invalid, without considering the number of tests administered. The proposed 30% rule accommodates this face validity question. It is tentatively proposed as a starting point for future research, and with sufficient empirical support, a general guideline for FDRA.
Academic collaboration recommendation (ACR) can help researchers find potential partners for research and thus promote academic innovation. Recent works mostly use graph learning-based methods to explore various ways of combining node information with topology, which consists of multiple steps, including network construction, node feature extraction, network representation learning and link prediction. One limitation is that they only conduct research on co-authorship networks and ignore citation relationship between publications. Besides, they tend to use attributes from a single dimension of researchers and do not take attributes of researchers from multiple dimensions into consideration at the same time. To address the above issues, we present the multi-dimensional attributes enhanced heterogeneous (MAH) network representation learning method, which constructs heterogeneous networks with both co-authorship and citation information and makes use of multi-dimensional attributes. Three research questions are addressed in this work: (RQ1) Is our proposed method effective on academic collaboration recommendation compared with existing state-of-the-art methods? (RQ2) Does incorporating citation information into co-authorship network help improve the performance of academic collaboration recommendation? (RQ3) How does fusing multi-dimensional attributes affect the performance of academic collaboration recommendation? A publicly available real-world data set is used in our experiments. The superior performance of MAH compared with baseline methods demonstrates that the proposed multi-dimensional feature-based researcher profile can enrich node information in academic network and effective researcher representations can be learned by applying graph representation learning methods on the network.
We demonstrate magnetic induction heating (MIH) with superparamagnetic iron oxide nanoparticles (IONPs) as a new rapid and energy-efficient methodology for synthesizing metal-organic frameworks (MOFs). Acting as localized heat sources, these...
Even though significant progress has been made in standardizing document layout analysis, complex layout documents like magazines and newspapers still present challenges. Models trained on standardized documents struggle with these complexities, and the high cost of annotating such documents limits dataset availability. To address this, we propose the Complex Layout Document Image Generation (DIG) model, which can generate diverse document images with complex layouts and authentic-looking text, aiding in layout analysis model training. Concretely, we first pre-train DIG on a large-scale document dataset with a text-sensitive loss function to address the issue of unreal generation of text regions. Then, we fine-tune it with a small number of documents with complex layouts to generate new images with the same layout. Additionally, we use a layout generation model to create new layouts, enhancing data diversity. Finally, we design a box-wise quality scoring function to filter out low-quality regions during layout analysis model training to enhance the effectiveness of using the generated images. Experimental results on the DSSE-200 and PRImA datasets show when incorporating generated images from DIG, the mAP of the layout analysis model is improved from 47.05 to 56.07 and from 53.80 to 62.26, respectively, which is a 19.17% and 15.72% enhancement compared to the baseline.
Legacies of an authoritarian past have enduring effects on voters’ attitudes and behaviors. I argue that authoritarian nostalgia is an important source of group sentiment and voter behavior in post-authoritarian democracies. Voters with nostalgic sentiment construct strong group sentiment based on historical perception and express attachment towards authoritarian successors. I test this argument with a new measure of authoritarian nostalgia. With original data collected from South Korea and Taiwan, I provide evidence that nostalgic voters are likely to exhibit strong group sentiment observable through partisan attachment. Abstracting from the specific cases, I use a randomly assigned candidate comparison analysis to demonstrate that voters high in authoritarian nostalgia are more attracted to hypothetical candidates invoking nostalgia than those with high programmatic or ideological proximity. Overall, the results show how authoritarian nostalgia remains important as a source of group sentiment in maturing democracies.
Turning is an important aspect of life underwater, playing integral roles in predator avoidance, prey capture, and communication. While turning abilities have been explored in a diversity of adult nekton, little is currently known about turning in early ontogeny, especially for cephalopods. In this study, we investigated the turning abilities of hatchling common cuttlefish (Sepia officinalis, n = 49) and dwarf cuttlefish (Sepia bandensis, n = 30), using both kinematic and wake-based analyses. Using body tracking software and particle image velocimetry (PIV), we found that S. officinalis turned faster than S. bandensis, but both species completed equally tight turns. Orientation (arms-first or tail-first) did not have a significant effect on turning performance for either species. Cuttlefish hatchlings used multiple short jets for more controlled turning, with jet mode I (isolated vortex rings) being 3–4 times more common than jet mode II (elongated jets with leading ring structures) for both species. While both hatchlings turned more broadly than adult squid and cuttlefish, S. officinalis hatchlings turned faster than adult cuttlefish, and both hatchlings turned more tightly than other jet-propelled animals and some non-jet-propelled swimmers.
Institution pages aggregate content on ResearchGate related to an institution. The members listed on this page have self-identified as being affiliated with this institution. Publications listed on this page were identified by our algorithms as relating to this institution. This page was not created or approved by the institution. If you represent an institution and have questions about these pages or wish to report inaccurate content, you can contact us here.
Information