In this research, we exploited the deep learning framework to differentiate the distinctive types of lesions and nodules in breast acquired with ultrasound imaging. A biopsy-proven benchmarking dataset was built from 5151 patients cases containing a total of 7408 ultrasound breast images, representative of semi-automatically segmented lesions associated with masses. The dataset comprised 4254 benign and 3154 malignant lesions. The developed method includes histogram equalization, image cropping and margin augmentation. The GoogLeNet convolutionary neural network was trained to the database to differentiate benign and malignant tumors. The networks were trained on the data with augmentation and the data without augmentation. Both of them showed an area under the curve of over 0.9. The networks showed an accuracy of about 0.9 (90%), a sensitivity of 0.86 and a specificity of 0.96. Although target regions of interest (ROIs) were selected by radiologists, meaning that radiologists still have to point out the location of the ROI, the classification of malignant lesions showed promising results. If this method is used by radiologists in clinical situations it can classify malignant lesions in a short time and support the diagnosis of radiologists in discriminating malignant lesions. Therefore, the proposed method can work in tandem with human radiologists to improve performance, which is a fundamental purpose of computer-aided diagnosis.

Deep metric learning aims to learn an embedding function, modeled as deep neural network. This embedding function usually puts semantically similar images close while dissimilar images far from each other in the learned embedding space. Recently, ensemble has been applied to deep metric learning to yield state-of-the-art results. As one important aspect of ensemble, the learners should be diverse in their feature embeddings. To this end, we propose an attention-based ensemble, which uses multiple attention masks, so that each learner can attend to different parts of the object. We also propose a divergence loss, which encourages diversity among the learners. The proposed method is applied to the standard benchmarks of deep metric learning and experimental results show that it outperforms the state-of-the-art methods by a significant margin on image retrieval tasks.

Abstract-We wish to minimize the resources used for network coding while achieving the desired throughput in a multicast scenario. We employ evolutionary approaches, based on a genetic algorithm, that avoid the computational complexity that makes the problem NP-hard. Our experiments show great improvements over the sub-optimal solutions of prior methods. Our new algorithms improve over our previously proposed algorithm in three ways. First, whereas the previous algorithm can be applied only to acyclic networks, our new method works also with networks with cycles. Second, we enrich the set of components used in the genetic algorithm, which improves the performance. Third, we develop a novel distributed framework. Combining distributed random network coding with our distributed optimization yields a network coding protocol where the resources used for coding are optimized in the setup phase by running our evolutionary algorithm at each node of the network. We demonstrate the effectiveness of our approach by carrying out simulations on a number of different sets of network topologies.

Although biophysical yield responses to local warming have been studied, we know little about how crop yield growth—a function of climate and technology—responds to global temperature and socioeconomic changes. Here, we present the yield growth of major crops under warming conditions from preindustrial levels as simulated by a global gridded crop model. The results revealed that global mean yields of maize and soybean will stagnate with warming even when agronomic adjustments are considered. This trend is consistent across socioeconomic assumptions. Low-income countries located at low latitudes will benefit from intensive mitigation and from associated limited warming trends (1.8 °C), thus preventing maize, soybean and wheat yield stagnation. Rice yields in these countries can improve under more aggressive warming trends. The yield growth of maize and soybean crops in high-income countries located at mid and high latitudes will stagnate, whereas that of rice and wheat will not. Our findings underpin the importance of ambitious climate mitigation targets for sustaining yield growth worldwide.

Background Collaboration is an ABET accreditation required component of the engineering curriculum. Research has shown that collaborative learning positively influences student achievement. The relationship between motivation, collaborative learning strategies, and achievement is not well studied in an engineering education context. Purpose(Hypothesis) A set of hypotheses were tested that predicted positive relationships between students' self‐reported informal collaboration, self‐efficacy for learning course material, knowledge building behaviors, and course grade. A second set of hypotheses were tested that predicted gender similarities in reported self‐efficacy, and gender differences in reported collaborative learning activities. Design/Method One hundred fifty engineering students were surveyed for study 1 and 513 students were surveyed for study 2. Bivariate correlations were completed to examine relationship between study variables; multiple regression analysis was completed to examine predictive ability of variables on course grade; MANOVA was completed to examine multivariate relationship between variables. Results In study 1, students' reported use of collaborative learning strategies and reported self‐efficacy for learning course material were significantly predictive of their course grade. In study 2, female students reported greater use of collaboration as a learning strategy than their male classmates; among male and female students combined, those who received “B's” in their engineering course reported more collaboration than their peers who received “A's” or “C's” and lower. Conclusion Overall, students' self reported collaborative learning strategies were associated with increased self‐efficacy for learning course material and course grade, particularly for students who received “B's” in the course. Female students reported greater use of collaborative learning strategies than their male peers.

Abstract. We demonstrate how a genetic algorithm solves the problem of minimizing the resources used for network coding, subject to a throughput constraint, in a multicast scenario. A genetic algorithm avoids the computational complexity that makes the problem NP-hard and, for our experiments, greatly improves on sub-optimal solutions of established methods. We compare two different genotype encodings, which tradeoff search space size with fitness landscape, as well as the associated genetic operators. Our finding favors a smaller encoding despite its fewer intermediate solutions and demonstrates the impact of the modularity enforced by genetic operators on the performance of the algorithm.

Figure 5. Operation of spin-encoded metaholograms in a broadband range of visible light. The metasurface optimized at 633 nm shows a negligible crosstalk between each image as shown in Figures 3 and 4. In other wavelengths, however, phase deviation (induced by propagation phase shift) causes crosstalk. Also, below 500 nm wavelength, extinction coefficient of a-Si:H is high enough to make the device lossy.

Partitioning ecosystem evapotranspiration (ET) into soil evaporation (E) and transpiration (T) is crucial for understanding hydrological processes. In this study, by using high-frequency isotope measurements and continuous surface water measurements, we investigated the isotope ratios in soil-vegetationatmosphere transfer and the physical mechanisms involved over a paddy field for a full growing season. The isotopic signals of d ET , d T , and d E were determined by the Keeling plot method, surface water isotopic measurements, and the Craig-Gordon model, respectively. The fraction of transpiration in evapotranspiration (FT) ranged from 0.2 to 1, with an almost continuous increase in the early growing season and a relatively constant value close to 1 later in the year. The result was supported by FT derived from simulated T and eddy correlation measured ET. The seasonal change in the transpiration fraction could be described quite well as a function of the LAI (FT 5 0.67LAI 0.25 , R 2 5 0.80), implying that transpiration plays a dominant role in the soil-vegetation-atmosphere continuum during the growing season. The two end-member uncertainty analysis suggested that further improvement in the estimation of d T and d ET is necessary for partitioning evapotranspiration using the isotopic method. In the estimation of d ET , the assumptions underlying Keeling plot method were rarely met and the uncertainty was quite large. A high frequency of precise isotopic measurements in surface water was also necessary for d T estimation. Furthermore, special care must be taken concerning the kinetic fractionation parameter in the Craig and Gordon Equation for d E estimation under low-LAI conditions. The results demonstrated the robustness of using isotope measurements for partitioning evapotranspiration.