Though semantic segmentation has been heavily explored in vision literature, unique challenges remain in the remote sensing domain. One such challenge is how to handle resolution mismatch between overhead imagery and ground-truth label sources, due to differences in ground sample distance. To illustrate this problem, we introduce a new dataset and use it to showcase weaknesses inherent in existing strategies that naively upsample the target label to match the image resolution. Instead, we present a method that is supervised using low-resolution labels (without upsampling), but takes advantage of an exemplar set of highresolution labels to guide the learning process. Our method incorporates region aggregation, adversarial learning, and self-supervised pretraining to generate fine-grained predictions, without requiring high-resolution annotations. Extensive experiments demonstrate the real-world applicability of our approach.

The enhancement of boiling heat transfer has been extensively shown to be achievable through surface texturing or fluid property modification, yet few studies have investigated the possibility of coupling both enhancement approaches. The present work focuses on exploring the possibility of concomitant enhancement of pool boiling heat transfer by using TiO2-water nanofluid in combination with laser-textured copper surfaces. Two mass concentrations of 0.001 wt.% and 0.1 wt.% are used, along with two nanoparticle sizes of 4–8 nm and 490 nm. Nanofluids are prepared using sonification and degassed distilled water, while the boiling experiments are performed at atmospheric pressure. The results demonstrate that the heat transfer coefficient (HTC) using nanofluids is deteriorated compared to using pure water on the reference and laser-textured surface. However, the critical heat flux (CHF) is significantly improved at 0.1 wt.% nanoparticle concentration. The buildup of a highly wettable TiO2 layer on the surface is identified as the main reason for the observed performance. Multiple subsequent boiling experiments using nanofluids on the same surface exhibited a notable shift in boiling curves and their instability at higher concentrations, which is attributable to growth of the nanoparticle layer on the surface. Overall, the combination of nanofluids boiling on a laser-textured surface proved to enhance the CHF after prolonged exposure to highly concentrated nanofluid, while the HTC was universally and significantly decreased in all cases.

Skin lesions can be an early indicator of a wide range of infectious and other diseases. The use of deep learning (DL) models to diagnose skin lesions has great potential in assisting clinicians with prescreening patients. However, these models often learn biases inherent in training data, which can lead to a performance gap in the diagnosis of people with light and/or dark skin tones. To the best of our knowledge, limited work has been done on identifying, let alone reducing, model bias in skin disease classification and segmentation. In this paper, we examine DL fairness and demonstrate the existence of bias in classification and segmentation models for subpopulations with darker skin tones compared to individuals with lighter skin tones, for specific diseases including Lyme, Tinea Corporis and Herpes Zoster. Then, we propose a novel preprocessing, data alteration method, called EdgeMixup, to improve model fairness with a linear combination of an input skin lesion image and a corresponding a predicted edge detection mask combined with color saturation alteration. For the task of skin disease classification, EdgeMixup outperforms much more complex competing methods such as adversarial approaches, achieving a 10.99% reduction in accuracy gap between light and dark skin tone samples, and resulting in 8.4% improved performance for an underrepresented subpopulation.

ABSTRACT Heat transfer coefficient (HTC) was experimentally measured for saturated and subcooled pool boiling of binary mixtures of water and glycerin. Saturated boiling was studied for mixtures with water mass fractions from to on horizontal flat nickel-plated surfaces at heat fluxes from 50 to at atmospheric pressure. Subcooled boiling was investigated in the range of subcooling from 0 to at heat fluxes of approximately 250, 450 and . It was found that mixture effects have a significant impact on saturated boiling HTC even for mixtures with very low content of glycerin as significant drops of HTC were observed for subtle changes in composition for mixtures of high . Measured HTC was successfully correlated with the combination of Yagov (1999) and Inoue and Monde (2009) correlations with a mean relative error of . A simple empirical HTC correlation is also proposed. For subcooled boiling, developed subcooled boiling regime was reached for all investigated heat fluxes. For this regime, correlations, which were able to predict HTC for saturated boiling, were employed to predict subcooled boiling HTCs for all investigated concentrations, heat fluxes and subcoolings. Effect of subcooling and effect of liquid composition on total HTC were of the same importance for mixtures with higher water content. With the increase in concentration of glycerin in the mixture, decrease in total HTC with increasing subcooling became more significant.

This article discusses how to create an interactive virtual training program at the intersection of neuroscience, robotics, and computer science for high school students with equity of access. A four-day microseminar, titled Swarming Powered by Neuroscience (SPN), was conducted virtually through a combination of presentations and interactive computer game simulations. The SPN microseminar was delivered by subject matter experts in neuroscience, mathematics, multi-agent swarm robotics, and education. The objective of this research was to determine if taking an interdisciplinary approach to high school education would enhance the students learning experiences in fields such as neuroscience, robotics, or computer science. This study found an improvement in student engagement for neuroscience by 16.6%, while interest in robotics and computer science improved respectively by 2.7% and 1.8%. The majority of students (64%) strongly agreed that they enjoyed learning from an interdisciplinary team of experts and 70% strongly agreed that the microseminar emphasized the need to have instruction teams with diverse disciplinary backgrounds. The curriculum materials, developed for the SPN microseminar, can be used by high school teachers to further evaluate interdisciplinary instructions across life and physical sciences and computer science.

Road transportation is one of the largest sectors of greenhouse gas (GHG) emissions affecting climate change. Tackling climate change as a global community will require new capabilities to measure and inventory road transport emissions. However, the large scale and distributed nature of vehicle emissions make this sector especially challenging for existing inventory methods. In this work, we develop machine learning models that use satellite imagery to perform indirect top-down estimation of road transport emissions. Our initial experiments focus on the United States, where a bottom-up inventory was available for training our models. We achieved a mean absolute error (MAE) of 39.5 kg CO2 of annual road transport emissions, calculated on a pixel-by-pixel (100 m2) basis in Sentinel-2 imagery. We also discuss key model assumptions and challenges that need to be addressed to develop models capable of generalizing to global geography. We believe this work is the first published approach for automated indirect top-down estimation of road transport sector emissions using visual imagery and represents a critical step towards scalable, global, near-real-time road transportation emissions inventories that are measured both independently and objectively.