Global Learning Strategy Research Articles

An architectural analysis of DeepOnet and a general extension of the physics-informed DeepOnet model on solving nonlinear parametric partial differential equations

The Deep Neural Operator, as proposed by Lu et al. (2021), marks a considerable advancement in solving parametric partial differential equations. This paper examines the DeepOnet model’s neural network design, focusing on the effectiveness of its trunk-branch structure in operator learning tasks. Three key advantages of the trunk-branch structure are identified: the global learning strategy, the independent operation of the trunk and branch networks, and the consistent representation of solutions. These features are especially beneficial for operator learning. Building upon these findings, we have evolved the traditional DeepOnet into a more general form from a network perspective, allowing a nonlinear interfere of the branch net on the trunk net than the linear combination limited by the conventional DeepOnet. The operator model also incorporates physical information for enhanced integration. In a series of experiments tackling partial differential equations, the extended DeepOnet consistently outperforms than the traditional DeepOnet, particularly in complex problems. Notably, the extended DeepOnet model shows substantial advancements in operator learning with nonlinear parametric partial differential equations and exhibits a remarkable capacity for reducing physics loss.

Multi-objective fuzzy partial disassembly line balancing considering preventive maintenance scenarios using enhanced hybrid artificial bee colony algorithm

In the existing researches about partial disassembly line balancing problem (PDLBP), all workstations are assumed always be available. In the real world, however, aging and failure of machines are inevitable, which may result in the shutdown of the whole disassembly line. To solve this problem, this study introduces preventive maintenance scenarios into PDLBP to prevent unexpected breakdowns and promote the productivity and smoothness of the disassembly line. Then, considering the substantial level of uncertainty in the state of end-of-life products, a multi-objective fuzzy mathematical model is established to minimize the cycle time, disassembly profit, and total assignment plan alteration simultaneously. And an enhanced hybrid artificial bee colony algorithm is developed to solve it. The attribute-driven heuristic strategy is applied to improve the quality of initial food sources, and the adaptive neighborhood search is designed to improve the exploitation efficiency of the employed bees and onlooker bees. Additionally, a hybrid global learning strategy is developed to guide the scout bees to raise the diversity of food sources. Finally, three case studies are conducted to validate the proposed algorithm. The results indicate that the proposed algorithm can achieve superior performance.

Constrained Planar Array Thinning Based on Discrete Particle Swarm Optimization with Hybrid Search Strategies.

This article presents a novel optimization algorithm for large array thinning. The algorithm is based on Discrete Particle Swarm Optimization (DPSO) integrated with some different search strategies. It utilizes a global learning strategy to improve the diversity of populations at the early stage of optimization. A dispersive solution set and the gravitational search algorithm are used during particle velocity updating. Then, a local search strategy is enabled in the later stage of optimization. The particle position is adaptively adjusted by the mutation probability, and its motion state is monitored by two observation parameters. The peak side-lobe level (PSLL) performance, effectiveness and robustness of the improved PSO algorithm are verified by several representative examples.

A novel hybrid fuzzy–metaheuristic approach for multimodal single and multi-objective optimization problems

In this paper, we propose a novel hybrid fuzzy–metaheuristic approach with the aim of overcoming premature convergence when solving multimodal single and multi-objective optimization problems. The metaheuristic algorithm used in our proposed approach is based on the imperialist competitive algorithm (ICA), a population-based method for optimization. The ICA divides its population into sub-populations, known as empires. Each empire is composed of a high fitness solution—the imperialist—and some lower fitness solutions—the colonies. Colonies move towards their associated imperialist to achieve better status (higher fitness). The most powerful empire tends to attract weaker colonies. These competitions and movements can be enhanced for better algorithm performance. In our hybrid approach, a global learning strategy is devised for each colony to learn from its best-known position, its associated imperialist and the global best imperialist. A fast-evolutionary elitism local search is used to enhance the collaborative search mechanism (competition) in each empire, and thus the overall optimization performance may be improved. Other main evolutionary operators include velocity adaptation and velocity divergence. To address parameterization and computational cost evaluation issues, two fuzzy inferencing mechanisms are designed and used in parallel: one is a learning strategy adaptor in each run, and the other is a smart evolution selector in each running window. For Pareto front approximation, fast-elitism non-dominated sorting is applied to the solutions, and a novel penalized sigma diversity index is designed to estimate the diversity (power) of solutions in the same rank. Comprehensive experimental results based on 22 single-objective and 25 multi-objective benchmark instances clearly show that our proposed approach provides better solutions compared with other popular metaheuristics and state-of-the-art ICA variants. The proposed approach can be used as an optimization module in any intelligent decision-making systems to enhance efficiency and accuracy. The designed fuzzy inferencing mechanisms can also be incorporated into any single- or multi-objective optimizers for parameter tuning purposes, to make the optimizers more adaptive to new problems or environments.

Developing the first-ever global learning strategy to frame the future of learning for achieving public health goals

Health is at the forefront of everyone's mind. Every country on the planet is currently experiencing the COVID19 pandemic, which is not just causing death and disease, but damaging societies and economies on a significant scale across the world.1 The current pandemic has brought into sharp focus the socio-economic factors and inequities in how people experience health and ill health and shone a spotlight on the need for lifelong learning so that health professionals, volunteers, and the public alike can adapt and respond to health threats.But even before the pandemic, the world was changing profoundly in the past few decades, and these changes impacted people's health. Unfortunately, the role of education and learning in the health domain has not kept pace with these changes. Many factors converge to create a major evolution, and some say, revolution, in how education, training and learning for health can and should be leveraged to protect and promote people's health. This paper summarizes some of the critical ideas of an 18-month process to create the first-ever global learning strategy for health by the World Health Organization.

A modified particle swarm optimization for multimodal multi-objective optimization

As an effective evolutionary algorithm, particle swarm optimization (PSO) has been widely used to solve single or multi-objective optimization problems. However, the performance of PSO in solving multi-objective problems is unsatisfactory, so a variety of PSO has been proposed to enhance the performance of PSO on multi-objective optimization problems. In this paper, a modified particle swarm optimization (AMPSO) is proposed to solve the multimodal multi-objective problems. Firstly, a dynamic neighborhood-based learning strategy is introduced to replace the global learning strategy, which enhances the diversity of the population. Meanwhile, to enhance the performance of PSO, the offering competition mechanism is utilized. 11 multimodal multi-objective optimization functions are utilized to verify the feasibility and effectiveness of the proposed AMPSO. Experimental results and statistical analysis indicate that AMPSO has competitive performance compared with 5 state-of-the-art multimodal multi-objective algorithms.

Integrating Sustainability into Higher Education Curricula through the Project Method, a Global Learning Strategy

Higher levels of material well-being lead almost inevitably to giving priority to individualism and personal advancement, often at the expense of civic conscience. A proposal for integrating sustainability into the curriculum is presented in the third year of the degree in Early Childhood Education at the Universitat Internacional de Catalunya (UIC). Projects on sustainable food are planned and elaborated to this aim. This study seeks to apply a global and systemic approach to solving socio-environmental problems and to check whether education for sustainable development (ESD) helps to develop and encourage actions that promote sustainable development. Quantitative research was conducted using a pre-test/post-test quasi experimental design separated by a period of didactic training in the project method. The results presented in this article show the students’ sustainability competencies (SC) improve after working on didactic proposals in a global manner. It is concluded that elaborating competencies in education for sustainable development enables an integrated approach of knowledge, procedures, attitudes and values in teaching through promoting the project method in multidisciplinary and transdisciplinary teams, which enhances future teachers’ sustainability competencies.

An efficient hybrid artificial bee colony algorithm for disassembly line balancing problem with sequence-dependent part removal times

ABSTRACTThis study deals with a sequence-dependent disassembly line balancing problem by considering the interactions among disassembly tasks, and a multi-objective mathematical model is established. Subsequently, a novel hybrid artificial bee colony algorithm is proposed to solve the problem. A new rule is used to initialize a bee colony population with certain diversity, and a dynamic neighbourhood search method is introduced to guide the employed/onlooker bees to promising regions. To rapidly leave the local optima, a global learning strategy is employed to explore higher quality solutions. In addition, a multi-stage evaluation method is designed for onlookers to effectively select employed bees to follow. The performance of the proposed algorithm is tested on a set of benchmark instances and two case scenarios, and the results are compared with several other metaheuristics in terms of solution quality and computation time. The comparisons demonstrate that the proposed algorithm exhibits superior performance.

Improved harmony search with general iteration models for engineering design optimization problems

Harmony search (HS) algorithm has a strong exploration and exploitation capability based on its unique improvisation. However, little research has been done on its improvisation mechanism. This paper offers a detailed discussion on the HS improvisation mechanism, which aims to state that the improvisation is a generic search framework. To improve the performance of HS, global learning strategy is designed to enhance the global search capability, and modified random selection is used to reduce the possibility of falling into local optimum. Moreover, a new improvement perspective such as the adjustment of iteration model is presented in this paper. Different iteration models such as dimension-to-dimension mode, stochastic multi-dimensional mode, vector mode and matrix mode to explore the optimization potential of HS algorithm are employed. Combining the improved operations, parameter adjustments and the four iteration models, four improved HS variants are proposed to analyze the effectiveness of iteration model on HS algorithm. Experimental results demonstrate the proposed HS algorithms can yield significant improved performance. Overall, the paper shows that the HS improvisation framework has a good extensibility and the iteration model has significant impact on the performance of HS.

A collaboration-based particle swarm optimizer for global optimization problems

This paper introduces a collaboration-based particle swarm optimizer (PSO) by incorporating three new strategies: a global learning strategy, a probability of learning, and a “worst replacement” swarm update rule. Instead of learning from the personal historical best position and the global (or local) best position which was used by the classical PSO, a target particle learns from another randomly chosen particle and the global best one in the swarm. Instead of accepting a new velocity directly, the velocity updates according to a learning probability, according to which the velocity of the target particle in each dimension updates via learning from other particles or simply inherits its previous velocity component. Since each particle has the same chance to be selected as a leader, the worst particle might influence the whole swarm’s performance. Therefore, the worst particle in the swarm in each update is moved to a new better position generated from another particle. The proposed algorithm is shown to be statistically significantly better than six other state-of-the-art PSO variants on 20 typical benchmark functions with three different dimensionalities.

RANKS: a flexible tool for node label ranking and classification in biological networks.

RANKS is a flexible software package that can be easily applied to any bioinformatics task formalizable as ranking of nodes with respect to a property given as a label, such as automated protein function prediction, gene disease prioritization and drug repositioning. To this end RANKS provides an efficient and easy-to-use implementation of kernelized score functions, a semi-supervised algorithmic scheme embedding both local and global learning strategies for the analysis of biomolecular networks. To facilitate comparative assessment, baseline network-based methods, e.g. label propagation and random walk algorithms, have also been implemented. The package is available from CRAN: https://cran.r-project.org/ The package is written in R, except for the most computationally intensive functionalities which are implemented in C. valentini@di.unimi.it Supplementary data are available at Bioinformatics online.

Training the world: using archetypes to create a practical global learning strategy

Purpose – The purpose of this paper is to define a new methodology for designing corporate learning for a global audience and to provide a case study of that methodology in action. The Global Learning Archetypes approach adapts well-established cultural preference models and combines them with insightful learning models. The result is three primary Global Learning Archetypes and six secondary archetypes that allow training to be designed once and used around the world. Design/methodology/approach – The Global Learning Archetype approach was created by evaluating well-established global cultural preferences models, integrating them with a proprietary learning criteria model, and developing a model for rapidly and cost-effectively creating learning for multiple geographies. Additionally, a case study illustrates both the challenges and successes when implementing this model in a large global corporation. Findings – Most organizations create global learning either by creating content in their “home” location and then adapting it for other locations, or by distributing a single version of content and trusting local facilitators to provide context for it. The first method is expensive and time-consuming; the second method is risky and unreliable. The Global Archetype method provides for creating learning interactions that are appropriate for multiple geographies in a single effort. Practical implications – Most large organizations are global, and smaller organizations increasingly have a global footprint. According to Fortune Magazine, the Fortune Global 500 are headquartered in 37 different countries and do business in over 150 different countries. An Institute for the Future/Intuit study notes that by 2018, half of all US small businesses will be involved in international trade. CSA Research observes that businesses spend about US$31 billion a year on localization. A method for providing global learning in both an impactful and cost-effective way is clearly necessary. Originality/value – The Global Learning Archetypes method is comparatively new, but it draws from well-established and well-vetted content on worldwide cultural preferences and on effective learning criteria. As such, it is a valuable synthesis of the proven and the innovative. Far more than a conceptual model, the Global Archetypes have been used by some of the largest organizations in the world; a case study of one such implementation is provided in this paper.

An extensive analysis of disease-gene associations using network integration and fast kernel-based gene prioritization methods.

ObjectiveIn the context of “network medicine”, gene prioritization methods represent one of the main tools to discover candidate disease genes by exploiting the large amount of data covering different types of functional relationships between genes. Several works proposed to integrate multiple sources of data to improve disease gene prioritization, but to our knowledge no systematic studies focused on the quantitative evaluation of the impact of network integration on gene prioritization. In this paper, we aim at providing an extensive analysis of gene-disease associations not limited to genetic disorders, and a systematic comparison of different network integration methods for gene prioritization.Materials and methodsWe collected nine different functional networks representing different functional relationships between genes, and we combined them through both unweighted and weighted network integration methods. We then prioritized genes with respect to each of the considered 708 medical subject headings (MeSH) diseases by applying classical guilt-by-association, random walk and random walk with restart algorithms, and the recently proposed kernelized score functions.ResultsThe results obtained with classical random walk algorithms and the best single network achieved an average area under the curve (AUC) across the 708 MeSH diseases of about 0.82, while kernelized score functions and network integration boosted the average AUC to about 0.89. Weighted integration, by exploiting the different “informativeness” embedded in different functional networks, outperforms unweighted integration at 0.01 significance level, according to the Wilcoxon signed rank sum test. For each MeSH disease we provide the top-ranked unannotated candidate genes, available for further bio-medical investigation.ConclusionsNetwork integration is necessary to boost the performances of gene prioritization methods. Moreover the methods based on kernelized score functions can further enhance disease gene ranking results, by adopting both local and global learning strategies, able to exploit the overall topology of the network.

Network-Based Drug Ranking and Repositioning with Respect to DrugBank Therapeutic Categories

Drug repositioning is a challenging computational problem involving the integration of heterogeneous sources of biomolecular data and the design of label ranking algorithms able to exploit the overall topology of the underlying pharmacological network. In this context, we propose a novel semisupervised drug ranking problem: prioritizing drugs in integrated biochemical networks according to specific DrugBank therapeutic categories. Algorithms for drug repositioning usually perform the inference step into an inhomogeneous similarity space induced by the relationships existing between drugs and a second type of entity (e.g., disease, target, ligand set), thus making unfeasible a drug ranking within a homogeneous pharmacological space. To deal with this problem, we designed a general framework based on bipartite network projections by which homogeneous pharmacological networks can be constructed and integrated from heterogeneous and complementary sources of chemical, biomolecular and clinical information. Moreover, we present a novel algorithmic scheme based on kernelized score functions that adopts both local and global learning strategies to effectively rank drugs in the integrated pharmacological space using different network combination methods. Detailed experiments with more than 80 DrugBank therapeutic categories involving about 1,300 FDA-approved drugs show the effectiveness of the proposed approach.

Cancer module genes ranking using kernelized score functions

BackgroundCo-expression based Cancer Modules (CMs) are sets of genes that act in concert to carry out specific functions in different cancer types, and are constructed by exploiting gene expression profiles related to specific clinical conditions or expression signatures associated to specific processes altered in cancer. Unfortunately, genes involved in cancer are not always detectable using only expression signatures or co-expressed sets of genes, and in principle other types of functional interactions should be exploited to obtain a comprehensive picture of the molecular mechanisms underlying the onset and progression of cancer.ResultsWe propose a novel semi-supervised method to rank genes with respect to CMs using networks constructed from different sources of functional information, not limited to gene expression data. It exploits on the one hand local learning strategies through score functions that extend the guilt-by-association approach, and on the other hand global learning strategies through graph kernels embedded in the score functions, able to take into account the overall topology of the network. The proposed kernelized score functions compare favorably with other state-of-the-art semi-supervised machine learning methods for gene ranking in biological networks and scales well with the number of genes, thus allowing fast processing of very large gene networks.ConclusionsThe modular nature of kernelized score functions provides an algorithmic scheme from which different gene ranking algorithms can be derived, and the results show that using integrated functional networks we can successfully predict CMs defined mainly through expression signatures obtained from gene expression data profiling. A preliminary analysis of top ranked "false positive" genes shows that our approach could be in perspective applied to discover novel genes involved in the onset and progression of tumors related to specific CMs.

Global Education Access Utilizing Partnerships and Networked Global Learning Communities

Networked global learning communities build partnership programs between higher education institutions and high schools which allow students, teachers and professors to attend and work in college preparation programs located in countries outside of their native lands. These educational programs help to promote development of transnational policies and procedure reforms to provide access to universities in other countries, as well as provide exposure to global learning strategies, structures, and emerging technologies among teachers and educational leadership. Transnational High School-University Bridge programs also allow the student to adjust to a new culture and work to improve their second language skills, while potentially earning college credit in a dual credit, high school environment.

Autostructuration of fuzzy systems by rules sensitivity analysis

We present a destructive (pruning) method aiming at gradually finding the appropriate number of rules in the case of fuzzy models. A particular attention has been paid to Takagi–Sugeno fuzzy systems (TS) for the problem of functions approximation. The proposed system can be seen as a generalization of the conventional TS system and allows to evaluate the importance of one particular rule in the inference process. The advantage of our approach has been put in light on two well-known benchmarks related to the field of chaotic time series forecasting. We also study and compare possible local and global learning strategies for these systems in terms of readability and performance.

Applications of information storage matrix neural networks

Information storage matrices (ISM) have recently been introduced as artificial neural networks. They define a new parallel processing architecture in which the neural connection weights are efficiently trained through a global learning strategy. This eliminates a need for slowly converging iterative learning schemes found in most present day neural network paradigms. Consequently, the ISM neural networks are attractive for real-time and online applications. Several such opportunities are discussed in this presentation. First it is demonstrated how the Boolean logic is implemented, and how the basic digital computer operations are realized through an inherently analog ISM organization. The ISM neural network is then trained to perform a pattern recognition task. The specified feature is extracted from a binary input data string. In addition, the ISM networks are used in the data processing and process control applications. In the first instance, spectral components of the covariance matrix are obtained without having to solve the characteristic equation. In the second case, process performance is stabilized on-line with respect to the initial state selected. As a result of such diverse application potentials, the ISM neural network is emerging as a useful parallel computing processor.

Effectiveness of a Global Learning Strategy Practiced in Different Contexts on Primary and Transfer Self-Paced Motor Tasks

(Singer, 1986) on achievement in laboratory and simulated self-paced sport tasks were investigated. Forty undergraduates were randomly stratified according to gender into four treatment groups: (a) a strategy group that initially practiced the strategy while learning the laboratory task (SL), (b) a laboratory control group that began the experiment by learning the task without the strategy (CL), (c) a strategy group that initially applied the strategy to the learning of an applied sport task (SA), and (d) a control group that initially learned the sport task without the strategy (CA). Following the completion of 48 trials with the primary task, all groups performed 50 trials on a transfer task. ANOVAs indicated that both strategy groups performed significantly better than their respective control groups in the primary tasks. Results of the transfer task indicated that the SA group performed at the same level as the SL group but outperformed both control groups. It was concluded that the strategy facilitates achievement in laboratory as well as applied self-paced tasks.

The Generalizability Effectiveness of a Learning Strategy on Achievement in Related Closed Motor Skills

Abstract This study determined the effectiveness of a global learning strategy on the skill level attained in one closed primary task and two related ones. Eighty subjects were randomly assigned to five groups equal in number and sex: a content-dependent strategy group (CDS), a content-dependent strategy plus reminders group (CDSR), a content-independent strategy group (CIS), a content-independent strategy plus reminders group (CISR), or a control group. A content dependent strategy is one learned in reference to a specific task; in this case, the primary one. A content independent strategy is learned in the context of many example tasks. Subjects performed underhanded dart throwing as the primary task, jart throwing (similar in nature to dart throwing) as the directly related task, and soccer foul shooting as the slightly related task. The strategy used to test the hypotheses generated was the Singer 5-Step Strategy. ANOVAs indicated that all of the strategy learning conditions were significantly more ef...