Mutual Learning Model Research Articles

DV-PSO-Net: A novel deep mutual learning model with Heuristic search using Particle Swarm optimization for Mango leaf disease detection

DV-PSO-Net: A novel deep mutual learning model with Heuristic search using Particle Swarm optimization for Mango leaf disease detection

Protocol of the study: Multilevel community-based mental health intervention to address structural inequities and adverse disparate consequences of COVID-19 pandemic on Latinx Immigrants and African refugees.

The NIMH-funded Multilevel Community-Based Mental Health Intervention to Address Structural Inequities and Adverse Disparate Consequences of COVID-19 Pandemic on Latinx Immigrants and African Refugees study aims to advance the science of multilevel interventions to reduce the disparate, adverse mental health, behavioral, and socioeconomic consequences of the COVID-19 pandemic that are a result of complex interactions between underlying structural inequities and barriers to health care. The study tests three nested levels of intervention: 1) an efficacious 4-month advocacy and mutual learning model (Refugee and Immigrant Well-being Project, RIWP); 2) engagement with community-based organizations (CBOs); and 3) structural policy changes enacted in response to the pandemic. This community-based participatory research (CBPR) study builds on long-standing collaboration with five CBOs. By including 240 Latinx immigrants and 60 African refugees recruited from CBO partners who are randomly assigned to treatment-as-usual CBO involvement or the RIWP intervention and a comparison group comprised of a random sample of 300 Latinx immigrants, this mixed methods longitudinal waitlist control group design study with seven time points over 36 months tests the effectiveness of the RIWP intervention and engagement with CBOs to reduce psychological distress, daily stressors, and economic precarity and increase protective factors (social support, access to resources, English proficiency, cultural connectedness). The study also tests the ability of the RIWP intervention and engagement with CBOs to increase access to the direct benefits of structural interventions. This paper reports on the theoretical basis, design, qualitative and quantitative analysis plan, and power for the study.

Development of policy measures for diffusing human pro-environmental behavior in social networks—Computer simulation of a dynamic model of mutual learning

Pro-environmental behavior does not diffuse sufficiently in society. Is there a way to enhance the degree of people's pro-environmental behavior? This study aims to develop a dynamic model of mutual learning in social networks to simulate the diffusion of pro-environmental behavior and to search for promising policies for promoting it. This study considers two policy measures: enhancing pro-environmental behavior of target people and changing the learning patterns of target people. The people targeted for intervention are determined by random selection, selection in descending order of degree centrality, and selection in descending order of eigenvector centrality. Centralities measure an influence of a node on other nodes through a network, based on the number of direct or indirect links. An interesting finding is that changing individual learning patterns is much more effective for enhancing the degree of pro-environmental behavior in social networks than trying to directly enhance its degree. In addition, selection of target people based on the centralities is more influential in encouraging environmentally friendly behavior than random selection, particularly in the policy of changing learning patterns. Multiplier effects are also measured: the ratio of the net increase in the number of people who enhance their degree of pro-environmental behavior at the end of a certain number of time steps beyond business as usual to the number of people intervened. Multiplier effects are always positive when learning patterns are changed. Six potential approaches to changing learning patterns are discussed: persuasion, reputation, competition, awareness of economic returns, information provisioning, and education.

A novel Vision Zero leadership training model to support collaboration and strategic action planning

Introduction: While collaboration and cooperation are regarded as foundational to Vision Zero (VZ) and Safe Systems initiatives, there is little guidance on structuring VZ collaboration, conducting collaborative goal setting, and aligning tangible action across organizations. As part of a larger VZ mutual learning model, we developed a VZ Leadership Team Institute to support communities in collaborative VZ strategic planning and goal setting. The purpose of this paper is to describe the development and evaluation of the Institute, which can serve as a foundation for other initiatives seeking to move VZ planning and implementation forward in a collaborative, systems-aware manner.Methods: In June 2021, eight multi-disciplinary teams of 3–6 persons each (n = 42 participants) attended the Institute, representing leaders from communities of various sizes. Surveys were administered pre, immediately post, and 6 months following the Institute. We measured confidence in a range of skills (on a 5-point scale, 1: not confident to 5: very confident). Surveys also measured coalition collaboration pre-Institute and 6 months post-Institute (on a 4-point scale, 1: strongly disagree to 4: strongly agree).Results: The largest increases in confidence from pre- to immediately post-Institute were for collaboratively drafting objectives and actions for VZ goals (pre-mean: 2.6, SD: 0.9 to post-mean: 3.8, SD: 0.9); incorporating equity into goals (pre-mean: 2.8, SD: 1.0 to post-mean: 3.9, SD: 0.8); and knowing how to keep VZ planning and implementation efforts on track (pre-mean: 2.6, SD: 1.0 to post-mean: 3.7, SD: 0.7). For all measures, average confidence in skills decreased from immediately post-Institute to 6 months post-Institute, but remained greater than average scores pre-Institute. Several measures of coalition collaboration maintained high agreement across time, and mean agreement increased for reporting that the future direction of the coalition was clearly communicated to everyone (pre-mean: .6, SD: 0.8; 6 months post-mean: 3.1, SD: 0.4). However, average scores decreased for feeling like the coalition had adequate staffing (pre-mean: 3.0, SD: 0.6; 6 months post-mean: 2.3, SD: 0.5).Discussion: The Institute utilized innovative content, tools, and examples to support VZ coalitions’ collaborative and systems-aware planning and implementation processes. As communities work toward zero transportation deaths and serious injuries, providing effective support models to aid multidisciplinary planning and action around a Safe Systems approach will be important to accelerate progress toward a safer transportation system.

MGL: Mutual Graph Learning for Camouflaged Object Detection.

Camouflaged object detection, which aims to detect/segment the object(s) that blend in with their surrounding, remains challenging for deep models due to the intrinsic similarities between foreground objects and background surroundings. Ideally, an effective model should be capable of finding valuable clues from the given scene and integrating them into a joint learning framework to co-enhance the representation. Inspired by this observation, we propose a novel Mutual Graph Learning (MGL) model by shifting the conventional perspective of mutual learning from regular grids to graph domain. Specifically, an image is decoupled by MGL into two task-specific feature maps - one for finding the rough location of the target and the other for capturing its accurate boundary details. Then, the mutual benefits can be fully exploited by reasoning their high-order relations through graphs recurrently. It should be noted that our method is different from most mutual learning models that model all between-task interactions with the use of a shared function. To increase information interactions, MGL is built with typed functions for dealing with different complementary relations. To overcome the accuracy loss caused by interpolation to higher resolution and the computational redundancy resulting from recurrent learning, the S-MGL is equipped with a multi-source attention contextual recovery module, called R-MGL_v2, which uses the pixel feature information iteratively. Experiments on challenging datasets, including CHAMELEON, CAMO, COD10K, and NC4K demonstrate the effectiveness of our MGL with superior performance to existing state-of-the-art methods. The code can be found at https://github.com/fanyang587/MGL.

Corporate Hierarchy and Organizational Learning: Member Turnover, Code Change, and Innovation in the Multiunit Firm

This study examines how recombinant innovation is affected by member turnover and organizational learning within a corporate hierarchy. Prior work has overlooked the role of organizational structure in organizational learning, focusing instead on the knowledge provided by individual new hires or on the disruption caused by individual departures. We address this gap by applying March’s [March JG (1991) Exploration and exploitation in organizational learning. Organ. Sci. 2(1):71–87.] mutual learning model to a corporate hierarchy. In doing so, we theorize how the contributions of corporate staff to socializing new employees and to learning from the organizational code may differ from those of the organization’s subunit members. Empirically, we examine the learning effects of aggregate corporate and subunit arrivals and departures on novel recombinant innovation by subunits. Using 24 years of Motorola company directories, we construct membership turnover measures for corporate and subunit employees and exploit patent data to capture recombinant innovation. Our results suggest that, whereas the influx of new ideas through arrivals may be critical, breaking the pattern of inertial behavior through departures is more important for recombinant innovation. Corporate departures matter most for recombinant innovation, a result that reflects not only corporate staff’s slower individual learning from the organizational code but also its ability to update that code more quickly. In supplementary analyses, we find different effects for technical and nontechnical staff and internal and external arrivals, as well as demonstrate the mutual learning mechanism using internal corporate documents to capture code change. Our study has strong implications for theories of organizational learning, strategic human capital, organization design, and innovation. Supplemental Material: The online appendices are available at https://doi.org/10.1287/orsc.2022.1618 .

A Novel Method of Chinese Herbal Medicine Classification Based on Mutual Learning

Chinese herbal medicine classification is an important research task in intelligent medicine, which has been applied widely in the fields of smart medicinal material sorting and medicinal material recommendation. However, most current mainstream methods are semi-automatic, with low efficiency and poor performance. To tackle this problem, a novel Chinese herbal medicine classification method based on mutual learning has been proposed. Specifically, two small student networks are designed for collaborative learning, and each of them collects knowledge learned from the other one respectively. Consequently, student networks obtain rich and reliable features, which will further improve the performance of Chinese herbal medicinal classification. In order to validate the performance of the proposed model, a dataset with 100 Chinese herbal classes (about 10,000 samples) was utilized and extensive experiments were performed. Experimental results verify that the proposed method is superior to those of the latest models with equivalent or even fewer parameters, specifically, obtaining 3∼5.4% higher accuracy rate and 13∼37% lower loss. Moreover, the mutual learning model achieves 80.8% Chinese herbal medicine classification accuracy.

Voluntary Organizational Learning and Incentives to be a Slow Learner

Studies on organizational learning emphasize the positive effect of slow learning in the long-run. We argue that this result hinges on the selectivity within the organization. Building on March’s (1991) classic model of mutual learning, we design the computational model to study the variation of learning rate effect, based on the different learning conditions by involving selectivity of both organizational code and its members. Our results suggest that the positive effects of slow learning vary depending on the selectivity of each entity. Specifically, we find that benefit of slow learning materializes when organizational code learn selectively. We also test this model across time to address the incentive issues of individuals, regarding to being a slow learner. We find that both learning rate and selectivity of entities should be considered important to seek the optimal learning conditions.

Person Re-IDentification Based on Mutual Learning With Embedded Noise Block

Some person re-identification(Re-ID) algorithms based on deep learning utilizes a baseline as basis to modify, and add some strategies to achieve better performance. Different from the conventional methods, this work combines distillation with mutual learning to construct a person Re-ID model of mutual learning. In training, in view of the characteristics of metric learning, we introduce a mutual loss <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$L_{M2}$ </tex-math></inline-formula> in the mutual learning network, so as to better promote the student networks to mine complementary information. In order to overcome the coupling problem in mutual learning, we designed a lightweight noise block and embedded it into mutual learning, which greatly improves the complementarity between networks. It should be added that the improvement achieved on the poor baseline can’t strictly prove the effectiveness of the research, so this paper constructs a person Re-ID baseline with relatively good performance, which is used as the student networks in mutual learning. Experiments demonstrate that the proposed person Re-ID algorithm based on mutual learning with embedded noise block achieves competitive performance on the Market1501, DukeMTMC-ReID, and CUHK-03 datasets.

«Research-policy Dialogues» in the Implementation of Migrants’ Integration Policy: Structure and Models in Russia

In this article, the author analyzes the structure and models of research-policy dialogues and identifies the conceptual emphasis of these interactions in the decision-making process in the scope of migrants integration policy in the Russian Federation. State authorities at all levels are making significant efforts to develop research and knowledge accumulation. Research-policy dialogues are important for the effective and the efficient development of migrants integration policy. The relevance of research-policy dialogues is determined by the need to use knowledge for various purposes by the state authorities: to provide a conceptual basis for political decisions in the field of migration policy, to develop tools and indexes for measuring integration policy, or to monitor and to evaluate the implemented policy. The author of the article analyzed the institutional forms of the research-policy dialogues in the field of migrant integration policy in the Russian Federation, studied communication channels in the framework of the St. Petersburg International Labor Forum (2018-2020), and analyzed the knowledge production process in this case using the network approach. The author considered the main research topics in the field of migration policy based on the data from the Unified Information System of Government Procurement (2016-2020), which allowed studying the functions of knowledge and the directions of its use. The author of the article concluded that the request for the mutual learning model of research-policy dialogues is just beginning to emerge in the Russian Federation. The most frequent models of dialogues are bureaucratic and consultancy, which require the use of knowledge in order to ensure legitimizing and substantiating functions.

Knowledge Gains and Efficiency Losses: Member Turnover and Organizational Performance

In order for organizations to avoid the fate of becoming rigid and obsolete over time, they need to update their knowledge base in a changing environment. One potential tool for this is member turnover that enables bringing in new knowledge. In this study, we model the effect of member turnover on organizational knowledge over the organizational life course with March’s [March, J. G. 1991. Exploration and exploitation in organizational learning. Organization Science, 2: 71–87] model of mutual learning and extend the estimation to broader organizational performance by including also the negative aspects of turnover, the diminishing internal cohesion and the shortening member tenure. The results show that there is a trade-off between knowledge gains and efficiency losses in the member turnover process. The balance between the positive and negative consequences of the process is also context dependent and changes over the life trajectory of an organization. The present work contributes to employee mobility literature by elucidating the twofold consequences of member turnover and to the studies on organizational aging by taking a temporal view on the process.

New Perspectives on Exploration and Exploitation: When Learning Fast is Superior to Slow Learning

Many organizations employ algorithms that learn from their members and then shape the way these individuals learn. Nevertheless, decades of research on organizational learning suggests that imperfect learning algorithms could sustain suboptimal beliefs that trap organizations indefinitely. To study potential algorithmic learning traps, we solve the underexplored theoretical properties of the March 1991 mutual learning model and demonstrate the conditions under which individuals should trust learning algorithms' recommendations. Our results show that the received wisdom regarding the benefit of slow learning and diversity does not hold when algorithms cannot identity accurate beliefs but follow the majority. The presence of non- discerning or even manipulated algorithms suggests that individuals should learn fast instead of slow to reduce the chance that algorithms learn the wrong, misleading lessons that would otherwise diffuse and contaminate everyone. Our exploitation of the March model generates novel insights that are increasingly relevant, thus promoting the model's generalization and making its beauty more robust.

Exploration, exploitation, and variability: Competition for primacy revisited

This study examines how exploration and exploitation contribute to variability in organizational performance and how this variability influences competitions for primacy, that is, contexts in which the ability to generate exceptionally high levels of performance is a key success factor. The results of simulation analyses conducted using the NK model, the mutual learning model, and the multiarmed bandit model all show that while exploration introduces more internal variability in organizations, exploitation tends to increase the variability in performance between different organizations. Variability and uniqueness in relation to one’s competitors, brought about by a single-minded focus on exploitation, is found to be advantageous in a race to finish first. In contrast, in contexts where avoiding especially low performance is the key consideration, exploration is found to be relatively more beneficial. These results suggest a need for a more nuanced understanding of the relationships between exploration, exploitation, and organizational risk taking.

The Role of Noise in a Mutual Selection and Learning Model

In the organizational learning and search literature, conventional wisdom holds that a moderate degree of noise helps organizations improve their performance by introducing more variation. We note, however, that portrayals of most organizational structures in this line of inquiry are overly simplified. We contribute to this literature by suggesting that the beneficial role of moderate noise is dependent on its relational presence in a mutual selection and learning system. By extending March’s (1991) mutual learning model, we examine the role of noise in an organization as an adaptive system where both the organizational code and individuals can select and learn from others. We demonstrate that, while a small amount of noise on the learning stage is beneficial, noise on the selection stage monotonically decreases performance. Our results indicate a need for re-examination of the role of noise in various organizational settings and structures.

Service productivity as mutual learning

Purpose – The purpose of this paper is to analyse the mutual learning implications for service productivity of the characteristics of service and service production. Design/methodology/approach – This is a conceptual paper. The starting point is, first of all, that productivity as a management concept should help a firm to manage its economic profit, and secondly, that service organizations are open systems, where the customers participate as co-producers and are exposed to the firm’s production resources and processes. Unlike in manufacturing, to understand productivity in service organizations as a means of managing profit, cost effects and revenue effects of changes in the productions system cannot be separated. Due to the interaction between customers and the firm’s resources during service production, dialogical collaboration between them develops. This enables mutual learning. Findings – Given the social dynamics in service production processes, four learning processes that influence service productivity are identified. Two processes enhance the organizations’s internal efficiency (cost savings), and two enhance its external effectiveness (perceived quality, revenue generation); two are organization-driven, two are customer-driven. Research limitations/implications – The mutual learning model demonstrates how the service provider by learning from the dynamics of service encounters in many ways can manage the productivity of the organizations’s processes. It shows that learning enables improvement of service productivity through effects enhancing both internal efficiency and external effectiveness. Originality/value – In a productivity context, learning has not earlier been studied as a mutual learning phenomenon.

A Missing Piece of Mutual Learning Model of March (1991)

(ProQuest: ... denotes formulae omitted.)IntroductionWhen multiple learners (including the organization) simultaneously affect each other, March referred to the process as the ecology of (Takase, 1991, p. 60). The ecology model of of March featured a selection process of organizational routines under the influence of other learners. A learner acquires other learners' experiences codified as technology, codes, procedures, and routines (Levitt & March, 1988). Furthermore, the diffusion (Rogers, 1962) of learners' experiences and routines in the organization makes this model even more complex.Mathematical analysis is typically abandoned and replaced by computer simulation analysis due to the complexity of the model. (e.g., Levinthal & March, 1981; Lounamaa & March, 1987). Throughout the 1980s, March was the primary proponent of this type of research (Huber, 1991).1 Levinthal and March (1993) called the phenomenon of preferring exploitation to exploration a myopia of learning (Sato, 2012). March (1991) entitled Exploration and Exploitation in Organizational primarily developed and analyzed two computer simulation models: (A) a mutual model and (B) a ecology model.However, Takahashi (1998) mathematically analyzes model (B) and does not require a computer simulation. In model (B), a reference organization having the normal performance distribution with mean m and variance v2 competes with N organizations having a standard normal performance distribution with mean 0 and variance 1. Let u be a value with an upper probability of 1/(N + 1) in the standard normal distribution table (u » 0.44 for N = 2, u » 1.34 for N = 10, u » 2.33 for N = 100), Takahashi (1998) obtains the following equation:... (1)Let P* be the probability of the reference organization having the best performance within the group. March (1991) constructs the competitive equality by conducting 5,000 simulations for each value v2 from 0 to 2 in steps of 0.05 and plotting points (m, v2) where P* = 1/(n + 1).The three lines of N = 2, N = 10, and N = 100 illustrated in Figure 6 in March (1991, p. 82) cross the vertical axis of mean m at approximately 0.2, 0.8, and 1.7, respectively. However, using Equation (1) the mean must be, respectively, 0.44, 1.34, and 2.33 when variance is 0. More specifically, Figure 1 illustrates the parabolas that touch the vertical axis at these points. Takahashi (1998) is suspicious regarding the validity of March's (1991) simulation program at least at the end point 0 of domain of variance.As March's (1991) programs have not been published, this paper creates a computer simulation program for model (A) (Appendix A) in free software language R for statistical computing, and checks the validity of the program at end points of the domain wherein we can analyze mathematically without simulation. We then compensate for the missing piece of March's (1991) mutual model (A) to identify the true conclusion.Mutual Learning ModelMarch's (1991) model is vaguely formulated because of less use of mathematical notation. This paper fully uses Takahashi's (1998) notation and reformulates March's (1991) model mathematically.RealityThere is an external reality, which is called the state of nature in statistical decision theory. A reality is an m-dimensional vector, with each component having a value of 1 or -1....Assumption 1: Each component ri of a reality is given an initial value of 1 or -1 with a probability of 1/2.The beliefs of nmembers and an organizational codeEach of n members has a belief for each component of the reality at each time period. This belief is not a probability, whereas the belief of statistical decision theory represents a probability. For each component in an m-dimensional vector of the reality, each belief has a value of 1, 0, or - 1. …

相互学習モデルの本当の結論:経営学輪講 March (1991)

相互学習モデルの本当の結論:経営学輪講 March (1991)

Balancing exploration and exploitation in complex environments

PurposeThe objective of this study is to model and analyze the exploration‐exploitation dynamics of March's model of mutual learning in a complex environment. By enhancing the above mentioned model, the paper seeks to propose a new agent‐based model of mutual learning within an organization.Design/methodology/approachThe paper replicates March's model of simulating learning within an organization using an agent‐based simulation approach, and extends it by modelling the problem space as a fitness landscape using Kauffman's NK model technique.FindingsIt was found that it is impossible to find a right balance between exploration and exploitation using the communication structure of March's model.Practical implicationsThe proposed model could help create a virtual laboratory for experimenting organizations' behavior in a complex co‐evolving environment. This virtual laboratory may be used in the future to support the decision‐making process of managers and policy makers.Originality/valueDesigning the external environment as a fitness landscape helps in discovering what effect the environmental complexity has on the emerging balance between exploration and exploitation. It is the first study to design the environment of a model which analyzes the mutual learning between an organization and its members as a complex non‐linear space.

Education Model of Mutual Learning on Students

Education Model of Mutual Learning on Students

Um estudo de simulação computacional para a análise de perfis de aprendizagem organizacional

Este artigo apresenta um modelo de simulação computacional que estende o modelo de aprendizagem mútua proposto por March (1991) com base no modelo SECI de Nonaka e Takeuchi (1997). Inicialmente são apresentadas considerações sobre aprendizagem organizacional e o modelo de aprendizagem mútua. Em seguida o modelo de March é revisitado, e são propostas extensões. São simulados diferentes cenários de evolução do modelo proposto, considerando as extensões, e analisados os resultados com base no referencial teórico. Finalmente, este trabalho apresenta considerações sobre a interpretação dos resultados obtidos em diferentes ambientes organizacionais, objetivando discutir a relação entre perfis de organização e aprendizagem organizacional.