Search

A crucial question in active patient care is determining if a treatment is having the desired effect, especially when changes are subtle over short periods. We propose using inter-patient data to train models that can learn to detect these fine-grained changes within a single patient. Specifically, can a model trained on multi-patient scans predict subtle changes in an individual patient's scans? Recent years have seen increasing use of deep learning (DL) in predicting diseases using biomedical imaging, such as predicting COVID-19 severity using lung ultrasound (LUS) data. While extensive literature exists on successful applications of DL systems when well-annotated large-scale datasets are available, it is quite difficult to collect a large corpus of personalized datasets for an individual. In this work, we investigate the ability of recent computer vision models to learn fine-grained differences while being trained on data showing larger differences. We evaluate on an in-house LUS dataset and a public ADNI brain MRI dataset. We find that models pre-trained on clips from multiple patients can better predict fine-grained differences in scans from a single patient by employing contrastive learning., Comment: Under review at ISBI 2025 conference

We present MELODI, a novel memory architecture designed to efficiently process long documents using short context windows. The key principle behind MELODI is to represent short-term and long-term memory as a hierarchical compression scheme across both network layers and context windows. Specifically, the short-term memory is achieved through recurrent compression of context windows across multiple layers, ensuring smooth transitions between windows. In contrast, the long-term memory performs further compression within a single middle layer and aggregates information across context windows, effectively consolidating crucial information from the entire history. Compared to a strong baseline - the Memorizing Transformer employing dense attention over a large long-term memory (64K key-value pairs) - our method demonstrates superior performance on various long-context datasets while remarkably reducing the memory footprint by a factor of 8.

The goal of this study was to provide insights on the extent to which high school students receive support for college and career planning, the mechanisms through which they receive this information, and how this varies across students and settings using data from a multi-state, multi-study project focused on advising for postsecondary transitions. Conducted by a collaborative team from three organizations--MDRC, the Research Alliance for New York City Schools, and University of North Carolina at Greensboro's Early College Research Center--this study examines advising structures and practices across 17 high schools in North Carolina and New York City, representing diverse school types, including career and technical education (CTE)-focused schools as well as comprehensive schools offering CTE for elective credit. Findings indicate that advising practices were largely shaped by each school's academic focus and staff expectations. CTE-focused schools, for example, typically were dually focused on preparing students for postsecondary education and careers through CTE programs that included structured partnerships with colleges and employers. In contrast, comprehensive schools varied in their emphasis on college and career advising and the degree to which students were expected to participate in career exploration and preparation activities. Some comprehensive schools expected or encouraged higher levels of CTE participation, whereas others offered CTE without requiring students to participate or commit to a pathway. These schools also differed in how career advising was integrated into the overall approach to postsecondary planning, with some focusing primarily on college preparation and others helping students develop career knowledge and skills as part of a broader college and workforce readiness strategy. Additionally, the study highlights the important role that advisors and other school personnel, particularly CTE teachers, play in helping students link their interests with available school-supported preparation activities and underscores the benefits of personalized, proactive advising tailored to individual student needs. [This paper was created by the Early College Research Center, UNC Greensboro.]

This paper presents a synthesis of three collaboratively conducted studies exploring the relationship between career-focused advising and the postsecondary transition with an emphasis on students enrolled in Career and Technical Education (CTE). The studies included a survey administered to high school seniors in New York City schools, an impact study of career coaches in North Carolina, and a qualitative study looking at implementation of advising in New York City and North Carolina. Key themes coming from the three studies included: 1) postsecondary transition advising is a schoolwide phenomenon; 2) the content and intensity shifts over a student's time in high school; 3) advising tends to focus more on college with less attention paid to career opportunities except in settings with an explicit career emphasis; 4) students with more advantaged backgrounds tend to participate in advising at higher levels; 5) higher participation in advising is associated with an increase in CTE-related activities; 6) college-focused advising is associated with higher enrollment in four-year schools while career-focused advising is associated with higher enrollment in two-year and lower enrollment in four-year institutions; and 7) additional research on how advising outcomes differ by student characteristics is needed. This article also summarizes key methodological takeaways about doing research related to advising. [This paper was created by the Early College Research Center, UNC Greensboro.]

The spin frequencies of neutron stars in low-mass X-ray binaries may be limited by the emission of gravitational waves. A candidate for producing such steady emission is a mass asymmetry, or "mountain", sourced by temperature asymmetries in the star's crust. A number of studies have examined temperature-induced shifts in the crustal capture layers between one nuclear species and another to produce this asymmetry, with the presence of capture layers in the deep crust being needed to produce the required mass asymmetries. However, modern equation of state calculations cast doubt on the existence of such deep capture layers. Motivated by this, we investigated an alternative source of temperature dependence in the equation of state, coming from the pressure supplied by the solid crustal lattice itself. We show that temperature-induced perturbations in this pressure, while small, may be significant. We therefore advocate for more detailed calculations, self-consistently calculating both the temperature asymmetries, the perturbations in crustal lattice pressure, and the consequent mass asymmetries, to establish if this is a viable mechanism for explaining the observed distribution of low-mass X-ray binary spin frequencies. Furthermore, the crustal lattice pressure mechanism does not require accretion, extending the possibility for such thermoelastic mountains to include both accreting and isolated neutron stars., Comment: 10 pages, 5 figures. To be published in MNRAS

Chemical reaction network theory is a field of applied mathematics concerned with modeling chemical systems, and can be used in other contexts such as in systems biology to study cellular signaling pathways or epidemiology to study the effect of human interaction on the spread of disease. In this paper, we seek to understand a chemical reaction network's equilibrium points through the lens of algebraic geometry by computing the positive part of the steady-state variety defined by polynomial equations arising from the assumption of mass-action kinetics. We provide a systematic classification of positive steady-state varieties produced by 2-species, 2-reaction networks, grounded in combinatorial and algebraic properties. While some (restricted) techniques exist to fully understand the ideal defining the positive steady-state variety, this computation presents a significant challenge in general. Our classification theorems provide a simplification of previous criteria, and aim to provide a foundation for future analysis of larger networks.

Recent advancements in robotics, including applications like self-driving cars, unmanned systems, and medical robots, have had a significant impact on the job market. On one hand, big robotics companies offer training programs based on the job requirements. However, these training programs may not be as beneficial as general robotics programs offered by universities or community colleges. On the other hand, community colleges and universities face challenges with required resources, especially qualified instructors, to offer students advanced robotics education. Furthermore, the diverse backgrounds of undergraduate students present additional challenges. Some students bring extensive industry experiences, while others are newcomers to the field. To address these challenges, we propose a student-centered personalized learning framework for robotics. This framework allows a general instructor to teach undergraduate-level robotics courses by breaking down course topics into smaller components with well-defined topic dependencies, structured as a graph. This modular approach enables students to choose their learning path, catering to their unique preferences and pace. Moreover, our framework's flexibility allows for easy customization of teaching materials to meet the specific needs of host institutions. In addition to teaching materials, a frequently-asked-questions document would be prepared for a general instructor. If students' robotics questions cannot be answered by the instructor, the answers to these questions may be included in this document. For questions not covered in this document, we can gather and address them through collaboration with the robotics community and course content creators. Our user study results demonstrate the promise of this method in delivering undergraduate-level robotics education tailored to individual learning outcomes and preferences., Comment: 11 pages, 3 figures, 1 table, 2024 ASEE Conference

To aid in prediction of turbulent boundary layer flows over rough surfaces, a new model is proposed to estimate hydrodynamic roughness based solely on geometric surface information. The model is based on a fluid-mechanics motivated geometric parameter called the wind-shade factor. Sheltering is included using a rapid algorithm adapted from the landscape shadow literature, while local pressure drag is estimated using a piecewise potential flow approximation. Similarly to evaluating traditional surface parameters such as skewness or average slope magnitude, the wind-shade factor is purely geometric and can be evaluated efficiently from knowing the surface elevation map and the mean flow direction. The wind-shade roughness model is applied to over 100 different surfaces available in a public roughness database and some others, and the predicted sandgrain-roughness heights are compared to measured values. Effects of various model ingredients are analyzed, and transitionally rough surfaces are treated by adding a term representing the viscous stress component., Comment: Paper submitted to J. Fluid Mechanics for publication, currently under review. Contains Notebooks but links will work only once it is published

This paper explores the use of large language models (LLMs) to score and explain short-answer assessments in K-12 science. While existing methods can score more structured math and computer science assessments, they often do not provide explanations for the scores. Our study focuses on employing GPT-4 for automated assessment in middle school Earth Science, combining few-shot and active learning with chain-of-thought reasoning. Using a human-in-the-loop approach, we successfully score and provide meaningful explanations for formative assessment responses. A systematic analysis of our method's pros and cons sheds light on the potential for human-in-the-loop techniques to enhance automated grading for open-ended science assessments., Comment: In press at EAAI-24: The 14th Symposium on Educational Advances in Artificial Intelligence

In recent years, large language models (LLMs) have become incredibly popular, with ChatGPT for example being used by over a billion users. While these models exhibit remarkable language understanding and logical prowess, a notable challenge surfaces in the form of "hallucinations." This phenomenon results in LLMs outputting misinformation in a confident manner, which can lead to devastating consequences with such a large user base. However, we question the appropriateness of the term "hallucination" in LLMs, proposing a psychological taxonomy based on cognitive biases and other psychological phenomena. Our approach offers a more fine-grained understanding of this phenomenon, allowing for targeted solutions. By leveraging insights from how humans internally resolve similar challenges, we aim to develop strategies to mitigate LLM hallucinations. This interdisciplinary approach seeks to move beyond conventional terminology, providing a nuanced understanding and actionable pathways for improvement in LLM reliability.

The efficiency of phonon-mediated heat transport is limited by the intrinsic atomistic properties of materials, seemingly providing an upper limit to heat transfer in materials and across their interfaces. The typical speeds of conductive transport, which are inherently limited by the chemical bonds and atomic masses, dictate how quickly heat will move in solids. Given that phonon-polaritons, or coupled phonon-photon modes, can propagate at speeds approaching 1 percent of the speed of light - orders of magnitude faster than transport within a pure diffusive phonon conductor - we demonstrate that volume-confined, hyperbolic phonon-polariton(HPhP) modes supported by many biaxial polar crystals can couple energy across solid-solid interfaces at an order of magnitude higher rates than phonon-phonon conduction alone. Using pump-probe thermoreflectance with a mid-infrared, tunable, probe pulse with sub-picosecond resolution, we demonstrate remote and spectrally selective excitation of the HPhP modes in hexagonal boron nitride in response to radiative heating from a thermally emitting gold source. Our work demonstrates a new avenue for interfacial heat transfer based on broadband radiative coupling from a hot spot in a gold film to hBN HPhPs, independent of the broad spectral mismatch between the pump(visible) and probe(mid-IR) pulses employed. This methodology can be used to bypass the intrinsically limiting phonon-phonon conductive pathway, thus providing an alternative means of heat transfer across interfaces. Further, our time-resolved measurements of the temperature changes of the HPhP modes in hBN show that through polaritonic coupling, a material can transfer heat across and away from an interface at rates orders of magnitude faster than diffusive phonon speeds intrinsic to the material, thus demonstrating a pronounced thermal transport enhancement in hBN via phonon-polariton coupling.

The relentless pursuit of miniaturization and performance enhancement in electronic devices has led to a fundamental challenge in the field of circuit design and simulation: how to accurately account for the inherent stochastic nature of certain devices. While conventional deterministic models have served as indispensable tools for circuit designers, they fall short when it comes to capture the subtle yet critical variability exhibited by many electronic components. In this paper, we present an innovative approach that transcends the limitations of traditional modeling techniques by harnessing the power of machine learning, specifically Mixture Density Networks (MDNs), to faithfully represent and simulate the stochastic behavior of electronic devices. We demonstrate our approach to model heater cryotrons, where the model is able to capture the stochastic switching dynamics observed in the experiment. Our model shows 0.82% mean absolute error for switching probability. This paper marks a significant step forward in the quest for accurate and versatile compact models, poised to drive innovation in the realm of electronic circuits.

Big data applications are on the rise, and so is the number of data centers. The ever-increasing massive data pool needs to be periodically backed up in a secure environment. Moreover, a massive amount of securely backed-up data is required for training binary convolutional neural networks for image classification. XOR and XNOR operations are essential for large-scale data copy verification, encryption, and classification algorithms. The disproportionate speed of existing compute and memory units makes the von Neumann architecture inefficient to perform these Boolean operations. Compute-in-memory (CiM) has proved to be an optimum approach for such bulk computations. The existing CiM-based XOR/XNOR techniques either require multiple cycles for computing or add to the complexity of the fabrication process. Here, we propose a CMOS-based hardware topology for single-cycle in-memory XOR/XNOR operations. Our design provides at least 2 times improvement in the latency compared with other existing CMOS-compatible solutions. We verify the proposed system through circuit/system-level simulations and evaluate its robustness using a 5000-point Monte Carlo variation analysis. This all-CMOS design paves the way for practical implementation of CiM XOR/XNOR at scaled technology nodes., Comment: 12 pages, 6 figures, 1 table

Recent policy efforts have attempted to increase the number of dual enrollment courses offered within Career and Technical Education pathways and there is evidence to suggest that this practice is widespread. However, there is very little research on student participation in CTE dual enrollment and on its impacts. This study examines participation in the CTE dual enrollment pathway in North Carolina, finding that about 9% of North Carolina students participated in CTE dual enrollment courses in 11th or 12th grade and disparities in participation among subgroups were less than for college transfer dual enrollment courses. Using a quasi-experimental approach to examine the effect of the program, the study found that participation is CTE was positively associated with college credits earned in high school, graduation from high school, and overall enrollment in college within one year after high school. The study also examined results by subgroup.

In this paper, we explore combinatorial properties of the posets associated with Kohnert polynomials. In particular, we determine a sufficient condition guaranteeing when such ``Kohnert posets'' are bounded and two necessary conditions for when they are ranked. Moreover, we apply the aforementioned conditions to find complete characterizations of when Kohnert posets are bounded and when they are ranked in special cases, including those associated with Demazure characters.

This paper presents a work-in-progress on a learn-ing system that will provide robotics students with a personalized learning environment. This addresses both the scarcity of skilled robotics instructors, particularly in community colleges and the expensive demand for training equipment. The study of robotics at the college level represents a wide range of interests, experiences, and aims. This project works to provide students the flexibility to adapt their learning to their own goals and prior experience. We are developing a system to enable robotics instruction through a web-based interface that is compatible with less expensive hardware. Therefore, the free distribution of teaching materials will empower educators. This project has the potential to increase the number of robotics courses offered at both two- and four-year schools and universities. The course materials are being designed with small units and a hierarchical dependency tree in mind; students will be able to customize their course of study based on the robotics skills they have already mastered. We present an evaluation of a five module mini-course in robotics. Students indicated that they had a positive experience with the online content. They also scored the experience highly on relatedness, mastery, and autonomy perspectives, demonstrating strong motivation potential for this approach., Comment: 5 pages, 2 figures, conference

This study presents an experimental dataset documenting the evolution of a turbulent boundary layer downstream of a rough-to-smooth surface transition. To investigate the effect of upstream flow conditions, two groups of experiments are conducted. For the \emph{Group-Re} cases, a nominally constant viscous-scaled equivalent sand grain roughness $k_{s0}^+\approx160$ is maintained on the rough surface, while the friction Reynolds number $Re_{\tau 0}$ ranges from 7100 to 21000. For the \emph{Group-ks} cases, $Re_{\tau 0}\approx14000$ is maintained while $k_{s0}^+$ ranges from 111 to 228. The wall-shear stress on the downstream smooth surface is measured directly using oil-film interferometry to redress previously reported uncertainties in the skin-friction coefficient recovery trends. In the early development following the roughness transition, the flow in the internal layer is not in equilibrium with the wall-shear stress. This conflicts with the common practise of modelling the mean velocity profile as two log laws below and above the internal layer height, as first proposed by Elliott (\textit{Trans. Am. Geophys. Union}, vol. 39, 1958, pp 1048--1054). As a solution to this, the current data are used to model the recovering mean velocity semi-empirically by blending the corresponding rough-wall and smooth-wall profiles. The over-energised large-scale motions leave a strong footprint in the near-wall region of the energy spectrum, the frequency and magnitude of which exhibit dependence on $Re_{\tau 0}$ and $k_{s0}^+$ respectively. The energy distribution in near-wall small scales is mostly unaffected by the presence of the outer flow with rough-wall characteristics, which can be used as a surrogate measure to extract the local friction velocity.

This paper examines recovery of the wall-shear stress of a turbulent boundary layer that has undergone a sudden transition from a rough to a smooth surface. Early works of Antonia and Luxton questioned the reliability of standard smooth-wall methods to measure wall-shear stress in such conditions, and subsequent studies show significant disagreement depending on the approach used to determine the wall-shear stress downstream. Here we address this by utilising a collection of experimental databases at Re_\tau \approx 4100 that have access to both `direct' and `indirect' measures of the wall-shear stress to understand the recovery to equilibrium conditions to the new surface. Our results reveal that the viscous region (z^+\lesssim 4) recovers almost immediately to an equilibrium state with the new wall conditions, however, the buffer region and beyond takes several boundary layer thicknesses before recovering to equilibrium conditions, which is longer than previously thought. A unique direct numerical simulation database of a wall-bounded flow with a rough-to-smooth wall transition is employed to confirm these findings. In doing so, we present evidence that any estimate of the wall-shear stress from the mean velocity profile in the buffer region or further away from the wall tends to underestimate its magnitude in the near vicinity of the rough-to-smooth transition, and this is likely to be partly responsible for the large scatter of recovery lengths to equilibrium conditions reported in the literature. Our results also reveal that the smaller energetic scales in the near-wall region recover to an equilibrium state associated with the new wall conditions within one boundary layer thickness downstream of the transition, while the larger energetic scales exhibit an over-energised state for several boundary layer thicknesses downstream of the transition.