Search results

The purpose of this paper is to study heterogeneous transfer learning for activity recognition using heuristic search techniques. Many pervasive computing applications require information about the activities currently being performed, but activity recognition algorithms typically require substantial amounts of labeled training data for each setting. One solution to this problem is to leverage transfer learning techniques to reuse available labeled data in new situations.

This paper introduces three novel heterogeneous transfer learning techniques that reverse the typical transfer model and map the target feature space to the source feature space and apply them to activity recognition in a smart apartment. This paper evaluates the techniques on data from 18 different smart apartments located in an assisted-care facility and compares the results against several baselines.

The three transfer learning techniques are all able to outperform the baseline comparisons in several situations. Furthermore, the techniques are successfully used in an ensemble approach to achieve even higher levels of accuracy.

The techniques in this paper represent a considerable step forward in heterogeneous transfer learning by removing the need to rely on instance – instance or feature – feature co-occurrence data.

The study aims to cope with the problems confronted in the skin lesion datasets with less training data toward the classification of melanoma. The vital, challenging issue is the insufficiency of training data that occurred while classifying the lesions as melanoma and non-melanoma.

In this work, a transfer learning (TL) framework Transfer Constituent Support Vector Machine (TrCSVM) is designed for melanoma classification based on feature-based domain adaptation (FBDA) leveraging the support vector machine (SVM) and Transfer AdaBoost (TrAdaBoost). The working of the framework is twofold: at first, SVM is utilized for domain adaptation for learning much transferrable representation between source and target domain. In the first phase, for homogeneous domain adaptation, it augments features by transforming the data from source and target (different but related) domains in a shared-subspace. In the second phase, for heterogeneous domain adaptation, it leverages knowledge by augmenting features from source to target (different and not related) domains to a shared-subspace. Second, TrAdaBoost is utilized to adjust the weights of wrongly classified data in the newly generated source and target datasets.

The experimental results empirically prove the superiority of TrCSVM than the state-of-the-art TL methods on less-sized datasets with an accuracy of 98.82%.

Experiments are conducted on six skin lesion datasets and performance is compared based on accuracy, precision, sensitivity, and specificity. The effectiveness of TrCSVM is evaluated on ten other datasets towards testing its generalizing behavior. Its performance is also compared with two existing TL frameworks (TrResampling, TrAdaBoost) for the classification of melanoma.

The purpose of this paper is to assess the impacts on production of five fruit crops from 1961 to 2018 of energy use, CO₂ emissions, farming areas and the labor force in China.

This analysis applied the autoregressive distributed lag-bound testing (ARDL) approach, Granger causality method and Johansen co-integration test to predict long-term co-integration and relation between variables. Four machine learning methods are used for prediction of the accuracy of climate effect on fruit production.

The Johansen test findings have shown that the fruit crop growth, energy use, CO₂ emissions, harvested land and labor force have a long-term co-integration relation. The outcome of the long-term use of CO₂ emission and rural population has a negative influence on fruit crops. The energy consumption, harvested area, total fruit yield and agriculture labor force have a positive influence on six fruit crops. The long-run relationships reveal that a 1% increase in rural population and CO₂ will decrease fruit crop production by −0.59 and −1.97. The energy consumption, fruit harvested area, total fruit yield and agriculture labor force will increase fruit crop production by 0.17%, 1.52%, 1.80% and 4.33%, respectively. Furthermore, uni-directional causality is correlated with the growth of fruit crops and energy consumption. Also, the results indicate that the bi-directional causality impact varies from CO₂ emissions to agricultural areas to fruit crops.

This study also fills the literature gap in implementing ARDL for agricultural fruits of China, used machine learning methods to examine the impact of climate change and to explore this important issue.

The purpose of this paper is to study organizational learning from complex and heterogeneous experiences. According to March (2010), this kind of high intellect learning is difficult to accomplish because it requires deliberate investments in knowledge transfer and creation. Zollo and Winter (2002) emphasized how knowledge codification can facilitate this process, as long as it is “well-performed”. However, knowledge management scholars have yet to explore what is meant by well-performed codification and how to achieve it.

This paper addresses this gap and provides a conceptual analysis based on two related but previously disconnected research areas: organizational learning and knowledge management.

This paper contributes to the literature in three ways. First, a new understanding of different types of experiences and their effects on learning is proposed. Then the codification process using a critical realist paradigm to overcome the epistemological boundaries of knowledge versus knowing is discussed; in doing so, it is shown that codification can take different forms to be “well-performed”. Finally, appropriate codification strategies based on experience type are identified.

The abstraction-oriented codification outlined in this paper runs counter to the logic of concrete codification that dominates both theory and practice. Thus, going beyond the traditional debate on the degree of codification (i.e. should knowledge be fully codified or just partly codified), this paper introduced a new debate about the appropriate degree of abstraction.

This paper aims to enlighten stakeholders about critical success factors (CSFs) in developing intelligent autonomous systems (IASs) by integrating artificial intelligence (AI) with robotics. It suggests a prioritization hierarchy model for building sustainable ecosystem for developing IASs.

This paper is based on the existing literature and on the opinion of 15 experts. All the experts have minimum of eight years of experience in AI and related technologies. The CSF theory is used as a theoretical lens and total interpretative structure modelling (TISM) is used for the prioritization of CSFs.

Developing countries like India could leverage IASs and associated technologies for solving different societal problems. Policymakers need to develop basic policies regarding data collection, standardized hardware, skilled manpower, funding and start-up culture that can act as building blocks in undertaking sustainable ecosystem for developing IASs and implementing national AI strategy. Clear-cut regulations need to be in place for the proper functioning of the ecosystem. Any technology that can function properly in India has better chances of working at the global level considering the size of the population.

This paper had all its experts from India only, and that makes the limitation of this paper, as there is a possibility that some of the factors identified may not hold same significance in other countries.

Stakeholders will understand the critical factors that are important in developing sustainable ecosystem for IASs and what should be the possible order of activities corresponding to each CSF.

The paper is the first of its kind that has used the CSF theory and TISM methodology for the identification and prioritization of CSFs in developing IASs. Further, eight significant factors, that is, emerging economy multinational enterprises (EMNEs), governance, utility, manpower, capital, software, data and hardware, have come up as the most important factors in integrating AI with robotics in India.

Hundreds of thousands of deaths each year in the world are caused by breast cancer (BC). An early-stage diagnosis of this disease can positively reduce the morbidity and mortality rate by helping to select the most appropriate treatment options, especially by using histological BC images for the diagnosis.

The present study proposes and evaluates a novel approach which consists of 24 deep hybrid heterogenous ensembles that combine the strength of seven deep learning techniques (DenseNet 201, Inception V3, VGG16, VGG19, Inception-ResNet-V3, MobileNet V2 and ResNet 50) for feature extraction and four well-known classifiers (multi-layer perceptron, support vector machines, K-nearest neighbors and decision tree) by means of hard and weighted voting combination methods for histological classification of BC medical image. Furthermore, the best deep hybrid heterogenous ensembles were compared to the deep stacked ensembles to determine the best strategy to design the deep ensemble methods. The empirical evaluations used four classification performance criteria (accuracy, sensitivity, precision and F1-score), fivefold cross-validation, Scott–Knott (SK) statistical test and Borda count voting method. All empirical evaluations were assessed using four performance measures, including accuracy, precision, recall and F1-score, and were over the histological BreakHis public dataset with four magnification factors (40×, 100×, 200× and 400×). SK statistical test and Borda count were also used to cluster the designed techniques and rank the techniques belonging to the best SK cluster, respectively.

Results showed that the deep hybrid heterogenous ensembles outperformed both their singles and the deep stacked ensembles and reached the accuracy values of 96.3, 95.6, 96.3 and 94 per cent across the four magnification factors 40×, 100×, 200× and 400×, respectively.

The proposed deep hybrid heterogenous ensembles can be applied for the BC diagnosis to assist pathologists in reducing the missed diagnoses and proposing adequate treatments for the patients.

This chapter investigates the antecedents to the development of the three components of subsidiaries’ absorptive capacity (ACAP): recognition, assimilation and application of transferred knowledge in the context of the vertical flow of social and environmental accounting and reporting (SEAR) knowledge from the HQ to acquired subsidiaries. Our analysis is based on an embedded multiple case study of a UK-based MNC, informed by 44 semi-structured interviews and capitalising on agency theory and socialisation theory. Prior knowledge is not a sufficient explanation to the development of ACAP but it is also dependent on organisational mechanisms that will trigger the learning processes. Depending on the nature and degree of the social, control and integration mechanisms, the effects of prior stocks of knowledge on ACAP may vary. Our propositions only hold for one direction of knowledge transfer. The study is based on an embedded multiple case study in one sector which restricts its generalisation. It excludes the specific relationships between the three ACAP learning processes and the existence of feedback loops. Our findings suggest that the HQ’s mix of social, control and integration mechanisms should account for initial stocks of SEAR knowledge. The contribution lies in uncovering the interaction between heterogeneous levels of prior knowledge and organisational mechanisms deployed by the HQ fostering ACAP. We address emerging issues regarding the reification of the ACAP concept and highlight the potential of agency theory for informing studies on HQ-subsidiary relations.

To combat poverty in China's rural areas, Chinese government has established an unconditional cash transfer program known as the Rural Minimum Living Standard Guarantee (Rural Dibao) Program. Interestingly, despite the importance of education in breaking cycles of poverty, little is known about Rural Dibao's impact on rural children's education. This study investigates Rural Dibao's impact on rural children's learning outcomes by first examining targeting issues within the program, exploring a causal relationship between Rural Dibao and learning outcomes, and then exploring potential mechanisms and heterogeneous effects.

Fixed effects model and propensity score weighting method and data from China Family Panel Studies (CFPS) from the years 2010 and 2014 were used.

The results suggest that the Rural Dibao program suffers from high levels of targeting error, yet is still effective (i.e., program transfers generally still go to people in need). The fixed effects and propensity score weighting models find that program participation raises rural children's standardized test scores in CFPS Chinese-language and math tests. In investigating mechanisms, increased education expenditure seems to connect Rural Dibao participation to increased learning results. The heterogeneity analysis shows that poorer, non-eastern, not left behind, younger or male children benefit from the program (while others have no effect).

These findings suggest that Rural Dibao participation boosts rural children's learning, which could indicate a long-term anti-poverty effect, and that if the program can resolve targeting problems, this effect could be even greater.

There is a need for higher education to produce graduates who are motivated to transfer learning into the workplace. Motivated graduates are work-ready and associated with increased performance. Presently, the research field around motivation to transfer learning by students in higher education is not clear and is inconsistent.

This scoping review provides an overview of the characteristics of the literature, including key concepts, recommendations and gaps based on eight published articles on the motivation of students in higher education to transfer learning.

The results reflected a research field, which focused primarily on the influence of specific factors, namely student characteristics, educational design, the workplace environment, and on higher education students' motivation to transfer learning. The lack of a shared conceptual definition of motivation to transfer learning in higher education appears to influence the description of the results from the included studies. Most of the previous studies applied rigorous research designs.

This seemingly stunted research field related to higher education students' motivation to transfer learning needs to be amplified to influence the development of work-ready graduates from higher education. Approaches towards including all elements of motivation, expanding to other fields in higher education, including low-income countries, may be a proximal step in enhancing the trajectory of this research field.