Search for: All records

Accurate streamflow prediction is critical for ensuring water supply and detecting floods, while also providing essential hydrological inputs for other scientific models in fields such as climate and agriculture.Recently, deep learning models have been shown to achieve state-of-the-art regionalization performance by building a global hydrologic model. These models predict streamflow given catchment physical characteristics and weather forcing data.However, these models are only focused on gauged basins and cannot adapt to ungaugaed basins, i.e., basins without training data. Prediction in Ungauged Basins (PUB) is considered one of the most important challenges in hydrology, as most basins in the United States and around the world have no observations. In this work, we propose a meta-transfer learning approach by enhancing imperfect physics equations that facilitate model adaptation. Intuitively, physical equations can often be used to regularize deep learning models to achieve robust regionalization performance under gauged scenarios, but they can be inaccurate due to the simplified representation of physics. We correct such uncertainty in physical equation by residual approximation and let these corrected equations guide the model training process. We evaluated the proposed method for predicting daily streamflow on the catchment attributes and meteorology for large-sample studies (CAMELS) dataset. The experiment results on hydrological data over 19 years demonstrate the effectiveness of the proposed method in ungauged scenarios. 

Inquiry-based course experiences provide a scalable and equitable way to engage students in research. In this study, we describe how we introduced inquiry-based experiences to ten lower-division and upper-division courses across the biology curriculum at a regionally comprehensive public university serving the diverse population in a major metropolitan area. Student survey data suggest this redesign effectively developed students’ scientific skills and nurtured their sense of belonging. This project illustrates how inquiry-based experiences can be implemented sustainably across institutional context. 

Active learning is a decision-making process. In both abstract and physical settings, active learning demands both analysis and action. This is a review of active learning in robotics, focusing on methods amenable to the demands of embodied learning systems. Robots must be able to learn efficiently and flexibly through continuous online deployment. This poses a distinct set of control-oriented challenges—one must choose suitable measures as objectives, synthesize real- time control, and produce analyses that guarantee performance and safety with limited knowledge of the environment or robot itself. In this work, we survey the fundamental components of robotic active learning systems. We discuss classes of learning tasks that robots typically encounter, measures with which they gauge the information content of observations, and algorithms for generating action plans. Moreover, we provide a variety of examples—from environmental mapping to nonparametric shape estimation—that highlight the qualitative differences between learning tasks, information measures, and control techniques. We conclude with a discussion of control-oriented open challenges, including safety-constrained learning and distributed learning. 

The impact of infiltrating chloride salts on the electrochemical behavior of lanthanum strontium manganite-yttria stabilized zirconia (LSM-YSZ) cathodes was investigated under solid oxide fuel cell operation. Infiltrating a lanthanum chloride solution resulted in the formation of a lanthanum oxychloride (LaOCl) phase. A LaOCl phase also formed by infiltrating an ammonium chloride solution; however, lanthanum was scavenged from the LSM phase to form LaOCl. The third infiltrating solution, a combination of zirconium chloride and yttrium nitrate, formed LaOCl by scavenging lanthanum from LSM and produced YSZ nanoparticles. Electrochemical impedance spectroscopy results suggest that LaOCl improves oxygen adsorption kinetics compared to a baseline LSM-YSZ cathode, reducing the low frequency impedance by 30%. In addition, scavenging lanthanum from LSM improved oxygen ion diffusion polarization as indicated by the observed 40% reduction in high frequency impedance and improved serial ohmic resistance by 19%. Finally, YSZ nanoparticles further reduced the high frequency impedance and ohmic resistance by 45% and 23%, respectively. The findings reveal new strategies for lowering the impedance of LSM-YSZ cathodes. 

The thermochemical stability of lanthanum strontium cobalt ferrite (LSCF) processed between 1000 °C–1200 °C via the in situ carbon templating method was studied. This method generates high surface area ceramics at traditional solid oxide fuel cell (SOFC) sintering temperatures by generating a carbon template in situ and subsequently removing the template by oxidation at 700 °C. Argon processed samples produced an amorphous carbon template, whereas nitrogen tended to form graphitic carbon. Prior to the oxidation step, nitrogen samples comprised larger La 2 O 3 crystallites (22–40 nm) compared to argon (9–17 nm). Upon oxidation, argon samples resulted in a pure LSCF phase with surface areas in the 21–29 m 2 ·g −1 range, whereas nitrogen samples contained significant impurities. This demonstrates that the size of La 2 O 3 crystallites formed during inert processing limited the ability to produce a pure LSCF phase. Symmetrical cells comprising nano-LSCF electrodes generated by the templating method were compared to cells sintered directly in air. Impedance results suggest that nano-LSCF cells and cells processed in air were dominated by interfacial charge transfer resistance and gas diffusion, respectively. The results map out conditions for preparing and integrating high surface area, nanostructured LSCF into SOFC electrodes at traditional sintering temperatures. Strategies for improving the interfacial resistance of nano-LSCF electrodes are discussed.