An artificial neural network (ANN) that mimics the information processing mechanisms and procedures of neurons in human brains has achieved a great success in many fields, e.g., classification, ...prediction, and control. However, traditional ANNs suffer from many problems, such as the hard understanding problem, the slow and difficult training problems, and the difficulty to scale them up. These problems motivate us to develop a new dendritic neuron model (DNM) by considering the nonlinearity of synapses, not only for a better understanding of a biological neuronal system, but also for providing a more useful method for solving practical problems. To achieve its better performance for solving problems, six learning algorithms including biogeography-based optimization, particle swarm optimization, genetic algorithm, ant colony optimization, evolutionary strategy, and population-based incremental learning are for the first time used to train it. The best combination of its user-defined parameters has been systemically investigated by using the Taguchi's experimental design method. The experiments on 14 different problems involving classification, approximation, and prediction are conducted by using a multilayer perceptron and the proposed DNM. The results suggest that the proposed learning algorithms are effective and promising for training DNM and thus make DNM more powerful in solving classification, approximation, and prediction problems.
Deep learning techniques have provided significant improvements in hyperspectral image (HSI) classification. The current deep learning-based HSI classifiers follow a patch-based learning framework by ...dividing the image into overlapping patches. As such, these methods are local learning methods, which have a high computational cost. In this article, a fast patch-free global learning (FPGA) framework is proposed for HSI classification. The proposed framework consists of three main parts: 1) a designed sampling strategy; 2) an encoder-decoder-based fully convolutional network (FCN); and 3) lateral connections between the encoder and decoder. In FPGA, an encoder-decoder-based FCN is utilized to consider the global spatial information by processing the whole image, which results in fast inference. However, it is difficult to directly utilize the encoder-decoder-based FCN for HSI classification as it always fails to converge due to the insufficiently diverse gradients caused by the limited training samples. To solve the divergence problem and maintain the FCNs abilities of fast inference and global spatial information mining, a global stochastic stratified (GS 2 ) sampling strategy is first proposed by transforming all the training samples into a stochastic sequence of stratified samples. This strategy can obtain diverse gradients to guarantee the convergence of the FCN in the FPGA framework. For a better design of FCN architecture, FreeNet, which is a fully end-to-end network for HSI classification, is proposed to maximize the exploitation of the global spatial information and boost the performance via a spectral attention-based encoder and a lightweight decoder. A lateral connection module is also designed to connect the encoder and decoder, fusing the spatial details in the encoder and the semantic features in the decoder. The experimental results obtained using three public benchmark data sets suggest that the FPGA framework is superior to the patch-based framework in both speed and accuracy for HSI classification.
Despite continued efforts by educators, UN declarations and numerous international agreements, progress is still limited in handling major global challenges such as ecosystem collapse, accelerating ...climate change, poverty, and inequity. The capacity to collaborate globally on addressing these issues remains weak. This historical review of research on global learning for sustainable development (GLSD) aims to clarify the diverse directions that research on GLSD has taken, to present the historical development of the research area, and highlight emerging research issues. The review summarizes key findings of 53 peer-reviewed publications, published in English in the period 1994–2020 identified with the search terms “global learning” and “sustainable development”, sustainability or GLSD, respectively. The review documented a gradually growing knowledge base, mostly authored by scholars located in the global North. Conclusions point to what we might achieve if we could learn from one another in new ways, moving beyond Northern-centric paradigms. It is also time to re-evaluate core assumptions that underlie education for sustainable development more generally, such as a narrow focus on formal learning institutions. The review provides a benchmark for future reviews of research on GLSD, reveals the emerging transformative structure of this transdisciplinary field, and offers reference points for further research.
This article explores Nordic countries’ and companies’ sustainability practices. It explores how nations like Denmark, Finland, and Sweden and companies such as Novo Nordisk and Ørsted achieve top ...sustainability rankings through their distinctive approach to stakeholder cooperation. It discusses the historical and cultural context that has shaped the Nordic approach, emphasizing the importance of long-term vision, stakeholder engagement, and cooperative strategies. It provides insights into how these practices contribute to achieving sustainable development goals and offers valuable lessons for global businesses seeking to integrate sustainable practices into their operations.
The work proposes a new approach for evolutionary global optimization which, in simple terms, may be described as a fusion of quantum and simulated annealing. This becomes possible by using basic ...concepts from Homotopy Theory, conjugated to the already existing Fuzzy Adaptive Simulated Annealing paradigm and other components. In this fashion, it occurs a temporal and nonlinear superposition of the two types of annealing, provoking interesting effects in runtime, including the so-called quantum tunneling, when there is a sudden transition between attraction basins of different minima without “climbing” potential barriers. In addition, the proposal includes a generalization of the “deformation” of Hamiltonians used in quantum annealing by suggesting the use of general homotopies between the surface corresponding to the cost function under processing and another specific landscape, being this possible because of the ability of Fuzzy ASA of handling time-varying objective functions. The proposed paradigm shows that it is possible and beneficial to superpose the two types of annealing, displaying amazing numerical results and suggesting a kind of interlacing or synergic behavior - perhaps we could call it an “annealing entanglement”. Finally, given the fast advances in commercial quantum annealing processors, it seems sensible to expect that an implementation of the proposed approach using such devices may occur very soon. In order to demonstrate the efficacy of the proposed algorithm, some significant and detailed examples are included in the text, so as to illustrate and clarify the presented ideas.
This paper examines whether, to what extent, and how international large-scale assessments (ILSAs) have influenced education policy-making at the national level. Based on an exploratory review of the ...research and policy literature on ILSAs and two surveys administered to educational policy experts, researchers, policymakers, and educators, our research found that ILSAs, with their multiple and ambiguous uses, increasingly function as solutions in search for the right problem - that is, they appear to be used as tools to legitimize educational reforms. The survey results pointed to a growing perception among stakeholders that ILSAs are having an effect on national educational policies, with 38% of respondents stating that ILSAs were generally misused in national policy contexts. However, while the ILSA literature indicates that these assessments are having some influence, there is little evidence that any positive or negative causal relationship exists between ILSA participation and the implementation of education reforms. Perhaps the most significant change associated with the use of ILSAs in the literature reviewed is the way in which new conditions for educational comparison have been made possible at the national, regional, and global levels.
Medical hyperspectral imaging provides new possibilities for non-invasive detection and characterization of diseases, and the processing of images can be accelerated and rationalized by using deep ...learning technology to classify pixels as one tissue or another, or as lesion or healthy tissue. However, most current methods for intelligently identifying pixels are not robust to large variations in pixel intensity within an image, particularly local learning approaches that rely on pixel or patch input. In this paper, we propose a network being able to learn to classify all pixels on an image by training with only a small number of manually labeled pixels in the same image. The network contains a hard band attention module (HBAM) to eliminate noisy bands and a dual-kernel spatial–spectral fusion attention module (DK-SSFAM) which uses two convolution kernels to weight spatial and spectral features and integrates them accordingly. We demonstrate that our proposed weakly supervised single-image global learning (SiGL) network classifies pixels in hyperspectral images of human brain in vivo better than traditional deep learning methods, suggesting potential for the clinic.
•The Global Teaching Model is situated locally, integrated with standards, critical framing, and intercultural collaboration.•Our data showed that global education was not a part of participants’ ...current curriculum.•However, teachers believed that global competence was important for them and their students.•Our findings show that teachers need guidance in translating global education theory to practice.
Education leaders recommend that global competence–global citizenship mentality and knowledge development for global participation–be incorporated into school curricula. This mixed methods study examined teacher’s perceptions and self-reported practices of globally competent teaching. Data was collected from teachers taking a graduate education course infused with global learning. Results suggest teachers value and desire to enact globally competent teaching but need practical direction for classroom effectuation. Data manifest all four dimensions of the Global Teaching Model (i.e., situated relevant practice, integrated global learning, critical and cultural consciousness raising, and intercultural collaboration for transformative action) to differing degrees. This study provides evidence for the Global Teaching Model as a prospective framework and emphasizes the critical dimension when internationalizing teacher education.
•Small class-imbalanced samples can inhibit the classification performance of the model.•A deep supervision global learning network with pair-weighted loss is designed to achieve a stronger ...prediction on small class-imbalanced datasets.•A semi-supervised learning method with confidence pseudo labels is proposed for increasing the diversity of samples.•Multi-scale prediction can improve classification performance.
Deep learning has been extensively applied in hyperspectral image (HSI) classification for its excellent representation ability. However, the existing training scheme generally adds a supervised classifier to the last layer of the network, so it is difficult to acquire full-scale fine-grained details and coarse-grained semantic information. Moreover, the robust performance of deep learning is commonly supported by numerous samples, so the effective discriminant features cannot be well learned with small class-imbalanced samples. To solve the above problems, a deeply-supervised pseudo learning framework (DSPL) is proposed, in which a deep supervision global learning network with pair-weighted loss is designed to achieve a stronger prediction on small class-imbalanced datasets, while this architecture of deep supervision can facilitate model generalization. To increase the diversity of samples, a semi-supervised learning method with confidence pseudo labels is proposed, capable of screening for more valid unlabeled samples and synthesizing some new mixed samples. To be more specific, the cost loss function consists of the supervised team (i.e., the labeled loss) and semi-supervised consistency regularization team (i.e., the unlabeled loss and the mixed loss), which can significantly enhance the generalization of the network by all useful samples. As revealed by the experimental results, the DSPL is better than other advanced methods on the Indian Pines (highest OA of 99.54% with 5% samples), the Pavia University (highest OA of 99.79% with 0.5% samples), as well as the Houston University 2013 (highest OA of 99.32% with 5% samples).
PDF
