Negative Transfer Research Articles

Valuable Beyond Vulnerable: A Scoping Review on the Contributions of Older Forced Migrants in Post-migration Recovery

AbstractDuring times of crisis, including pandemics, climate change, and forced migration, much of the discourse in ageing research and intervention centers on the vulnerabilities of older adults. Unfortunately, the valuable contributions of older adults to post-disaster recovery and healing are often overlooked and undervalued. Our aim in this scoping review is to shed light on the critical contributions of older forced migrants to post-migration recovery. We set the scene by introducing the two significant global demographic changes of the twenty-first century: forced migration and ageing. We provide a discourse on older forced migrants, ageing in situations of forced migration, and some of the challenges faced by older forced migrants. We then present some of the substantial roles of older forced migrants in post-migration recovery, including building resilience, contributing to culture and language transfer, providing emotional support, offering mentorship and leadership, participating in community building, and fostering social integration. We close by highlighting some of the lessons that can be drawn from understanding the unique roles played by older adults in post-forced migration recovery and the key actions necessary to promote these roles.

Beyond Affect Transfer: Attribute Associations in Video-Game Brand Placements and Their Impact on Brand Attitudes

Brand placements are omnipresent in video games. However, when and how they influence players’ brand attitudes is still insufficiently understood. Whereas previous research argued that positive affect transfers from the game to the placed brands and elevates brand attitudes, we propose and test that players form associations of the placed brands with specific attributes depending on the specific placement context. In a 2 × 2 within-subjects experiment with a third-person adventure game (N = 76), we strategically placed brands alongside positive versus negative in-game experiences by letting players collect items, conveying either speed or healthiness as an attribute. Brand attitudes became more positive if brands were paired with positive in-game experiences, irrespective of the specific attribute. However, brands also became associated with the distinct attributes they were paired with. These associations mediated the effects of the placement valence on brand attitudes, depending on which attribute was conveyed in the in-game experience. Our findings suggest that positive brand placement effects on brand attitudes go beyond a direct affect transfer but also arise from associations with specific attributes. Our study offers important insights into brand attitude formation from in-game advertising and valuable suggestions on effective brand placement.

Co-electrodeposition of superhydrophobic NiCo/CeO2 coating with hierarchical nano/micro structure for corrosion protection of plain carbon steel

This study aimed to develop an effective corrosion protection strategy for carbon steel by employing a superhydrophobic NiCo coating modified with CeO2 nanoparticles through the co-electrodeposition method. The superhydrophobicity was achieved by adjusting the composite coating's morphology and surface roughness. The synergistic impact of CeO2 and EDA crystal modifier on the coating's morphology, composition, topography, corrosion protection performance, and wetting properties was systematically investigated. The SEM results revealed the formation of a flower-like nano/micro morphology in the presence of EDA, and this morphology was preserved with the addition of 4 g/L CeO2 into the electrolyte. Surface wettability was significantly influenced by the hierarchical nano/micro structure of the deposited coating, with an enhanced superhydrophobicity observed in the presence of CeO2. The increase in surface roughness and compaction, achieved by incorporating CeO2 nanoparticles into the hierarchical structure of the NiCo coating, played a pivotal role in creating a superhydrophobic surface. Furthermore, polarization and EIS tests demonstrated that the superhydrophobic NiCo-EDA/CeO2 coating exhibited excellent anti-corrosion performance for the carbon steel substrate. In this case, Ecorr shifted towards positive values, icorr decreased, and both charge transfer resistance and coating resistance showed a considerable increase compared to other studied coatings. The alteration of metal wettability to a superhydrophobic state, combined with the barrier role of the coating and the presence of ceramic reinforcements (CeO2), offers comprehensive protection to the steel substrate. This approach holds great promise for effectively controlling the corrosion rate of widely used engineering metals.

Selecting reliable instances based on evidence theory for transfer learning

The aim of transfer learning is to improve the performance of learning models in the target domain by transferring knowledge from the related source domain. However, not all data instances in the source domain are reliable for the learning task in the target domain. Unreliable source-domain data may lead to negative transfer. To address this problem, we propose a novel strategy for selecting reliable data instances from the source domain based on evidence theory. Specifically, a mass function is formulated to measure the degree of ignorance and reliability of the source domain data with respect to the learning task in the target domain. By selecting reliable instances with low degree of ignorance from the source domain, the domain adaptation of the transfer learning models is enhanced. Moreover, the proposed data-selection strategy is independent of specific learning algorithms and can be regarded as a common preprocessing technique for transfer learning. Experiments on both simulated and real-world datasets validated that the proposed data selection strategy can improve the performance of various types of transfer learning methods.

Interlanguage Pragmatic Competence of University Students: An Error Analysis of Apology Speech Act Strategies in Japanese Learners

While acquiring a second language, learners may encounter challenges and difficulties in effectively carrying out verbal communication in the second language. Mastering the apology speech act is a challenge for L2 learners. The objective of this study is to identify the apology strategies utilized by individuals learning the Japanese language, as well as the specific types and underlying causes of errors produced during the execution of apologetic speech acts. The results of this study were gathered via the Discourse Completion Test (DCT), which encompassed a sample of 150 Japanese language learners. The gathered data were further classified using eight semantic formulas based on the research conducted by Haristiani and Sopiyanti (2019), while the various types and sources of error categorization were conducted based on the theories proposed by Corder (1981) and Richards (1975). The results of this study suggest that Japanese learners utilize the same primary strategies in apology speech acts. Furthermore, the study findings indicated that learners at the intermediate level displayed a higher frequency of errors in their speech compared to learners at the beginner levels and pre-intermediate levels, particularly in the strategy of taking responsibility. The main reason for this was primarily a lack of familiarity with the conventions of sentence structures, the proper use of language, and the appropriate use of expressions. The other sources of errors in language learning are likely to be ignoring language rule restrictions, incomplete application of rules, the learner’s false hypothesis, overgeneralization, and language transfer. This study is anticipated to function as a point of reference for research in interlanguage pragmatics, second-language acquisition, and error analysis. Gaining comprehension of the various problems and difficulties encountered when performing the speech act of apology in Japanese can assist both learners and educators in reducing these errors.

Semi-physical simulation-driven contrastive decoupling net for intelligent fault diagnosis of unseen machines under varying speed

Multi-source domain generalization provides a good idea for intelligent fault diagnosis of unseen machines. However, the acquisition of multi-source training sets would not only affect the generalization effect, but also cause huge economic loss. Too much difference between source training sets would lead to negative transfer, while too much similarity between the sets would cause zero transfer. Simultaneously, the current methods based on the assumption of stable speed cannot realize intelligent diagnosis under varying speed. Therefore, a semi-physical simulation-driven contrastive decoupling network is presented. First, a phenomenological model is used to generate large amount of simulation data with different modulation frequencies and signal-to-noise ratios to provide support for the training of the net. Next, with the joint help of simulation data and physical data, a contrastive framework is introduced to extract domain-independent features. Finally, a decoupling framework is proposed to minimize speed-related information while maximizing health-related features, thereby enabling intelligent diagnosis of unseen devices at varying speed. The proposed method is tested on two cases operating under sharp speed variation and compared with some state-of-the-art methods. The results show the superiority of the proposed net in intelligent diagnosis of unseen equipment under drastically varying speed.

The causes of negative countertransference in its cultural aspect among psychiatric residents in Tunisia

IntroductionNegative countertransference in psychiatry refers to the therapist’s unfavorable emotional reactions to the patient, such as anger and frustration, which can hinder the therapeutic relationship and the client’s progress. This is why it is imperative to study the causes of this negative counter-transference, such as cultural causes, to ensure effective treatment, appropriate care and better comfort for psychiatry residents during their professional practice.ObjectivesTo study the cultural causes of negative countertransference among psychiatric residents in Tunisia and their coping behavior.MethodsThis cross-sectional study was carried out among Tunisian residents working in psychiatric departments, using a questionnaire deployed via Google Forms.Results The study involved 26 residents with 23 females. The average age was 27.57 years with extremes ranging from 26 to 32. The participants were family doctors practicing in psychiatric wards (26.9%), first year psychiatry residents (15.4%), second year psychiatry residents (23.1%), third year psychiatry residents (19.2%), fourth year psychiatry residents (11.5%) and child psychiatry residents (3.8%). The majority of residents admitted having had a negative transference towards a patient (88.5%). The level of frustration felt by residents during this counter-transference on a scale of 100 varied from 1 to 100 with an average of 61.9. Substance abuse was the primary cause in 52.17% of cases. The second cause was the patient’s ideology, with a percentage equal to 43.47%. The same percentage of 17.39% was for traditions, socio-economic level and membership of a particular political group. In 82.6% of cases, residents tried to analyze this counter-transference and 65.2% of them managed to deal with their frustration. The feeling of guilt was experienced by 56.52% of practitioners and the same number of residents tried to avoid the patient. Among the participants, 43.47% discussed this difficulty with their supervising physician and only 4 residents asked to change patients.Conclusions In conclusion, our study identified the cultural causes of negative countertransference in Tunisian psychiatry residents, including substance abuse, ideology, traditions, socio-economic level and politics. Understanding these causes is essential to resident training but also to the delivery of quality care in psychiatry. By integrating this knowledge into training, we can help residents recognize and manage negative countertransference, in order to improve the quality of care they provide to their patients.Disclosure of InterestNone Declared

Challenges faced by English education in ethical minority areas under the background of Bilingual Education Policy for ethical minority students: Taking Yunnan Province as an example

With the promulgation of the national bilingual education policy, the vast majority of ethnic minorities have learned to master both Mandarin Chinese and ethnic minority languages, providing them with more opportunities for their development, but it does not emphasize their foreign language skills to go out into the world. Therefore, the research theme of this paper is the challenges of English language education in ethnic minority areas in the context of bilingual education, taking Yunnan Province as an example. The study finds that there are three main challenges facing English language education for ethnic minorities. They are language teaching issues in ethical minority areas, English teacher issues in ethical minority areas, and people’s attitudes toward English learning in ethical minority areas. Firstly, there are a lot of problems with English teachers, such as the impact of minority languages on Han English teachers is big, insufficient teacher resources, low professional quality, and high turnover. Then, the students learn English with language transfer and are not interested in it. Finally, schools, teachers, parents, and students do not value English education enough.

Autorotation Transfer of Training: Effects of Helicopter Dynamics

This paper analyzes the effects of the helicopter dynamics on pilots’ learning process and transfer of learned skills during autorotation training. A quasi-transfer-of-training experiment was performed with 10 experienced helicopter pilots in the SIMONA moving-base flight simulator at Delft University of Technology. Pilots had to control an in-house flight dynamics model setup to simulate two types of helicopter dynamics: (1) a “hard” dynamics characterized by a low autorotative flare index requiring high pilot control compensation and (2) a “easy” dynamics characterized by a high autorotative flare index with low pilot control compensation required. Two groups of pilots tested these types of dynamics in a different training sequence: hard-easy-hard (HEH group) and easy-hard-easy (EHE group). The main conclusion of this study proved that simulator training for autorotation can best start with pilots training in the most resource demanding condition. A more challenging helicopter’s dynamics will require higher pilot agility and more rapid responses to his/her perceptual changes. This will result in pilots developing more robust and adaptable flying skills. Indeed, a clear positive transfer of training effect was observed in the experiment presented in this paper in terms of acquired pilot skills in the HEH group, but not the EHE group. Positive transfer was especially observed in terms of reduced rate of descent at touchdown. The two groups differed in the control strategy applied, with the HEH group having developed a control technique mimicking more closely the one adopted in a real helicopter.

Mitigating L1 interference: Innovative tools and strategies

Mitigating L1 interference: Innovative tools and strategies

Multi-Agent Interactive Localization: A Positive Transfer Learning Perspective

Multi-Agent Interactive Localization: A Positive Transfer Learning Perspective

THE LANGUAGE INTERFERENCE IN ENGLISH CONVERSATION AT PONDOK PESANTREN FADHLUL FADHLAN SEMARANG DUE TO THE ADDITION OF JAVANESE PARTICLES

This research aimed to describe (1) the form and function of Javanese particles in English conversations; and explained the factors behind the occurrence of language interference through the addition of Javanese particles to English conversations by students at the Fadhlul Fadhlan Islamic boarding school in Semarang. This research included field research with a qualitative approach and was presented descriptively. The data for this research was in the form of oral speech from the data source of students at the Fadhlul Fadhlan Islamic Boarding School when they were in the Islamic boarding school environment. The data collection technique method used non-participant observation by observing students' conversations and then taking notes and structured interviews with 27 students. The data analysis technique method used three stages, namely data reduction, data presentation, and concluding. The interference theory by weinrich is to identify the interference phenomena. The results of this research showed that there were two types of Javanese particles added by students in English conversations, namely the single Javanese particles and the combination of Javanese particles. The second result was there were two forms of interference were found, namely grammatical interference and lexical interference. Third, the factors that caused interference through the addition of particles were internal factors in the form of bilingualism, lack of mastery of English grammar and vocabulary, and being carried away by mother tongue habits. Furthermore, external factors were the environment, facilitating communication between students, situations, and conditions

The Investigation into Pronunciation Problems Faced by Teacher-Trainees: A Case Study at the National Institute of Education (NIE) in Cambodia

This study investigates pronunciation challenges among English major -students at National Institute of Education, focusing on segmental and supra-segmental pronunciation issues, contributing factors, and potential remediation strategies. The study employed qualitative approach utilizing research apparatuses adjusted from previous studies. Additionally, ten teacher trainees with at least six years of English language learning background consented to participate in this study, while four lecturers were also incorporated. The findings were identified primarily with labiodental, dental, velar, alveolar, and palatal consonants, and with the monophthong sounds /i:/ and /ɒ/, and diphthongs /eɪ/ and /ɪə/. Factors influencing pronunciation proficiency include inadequate qualified teachers, insufficiently structured language programs, first language interference, and limited access to learning resources, alongside motivational and environmental factors, age, and beliefs. Recommendations for improvement emphasize integrating technology and social media for accessing pronunciation tutorials, extending instructional hours, and encouraging regular English-speaking practices.

Investigating Syntactic Transfer from English to Korean in Neural L2 Language Models

This paper investigates how the grammatical knowledge obtained in the initial language (English) of neural language models (LMs) influences the learning of grammatical structures in their second language (Korean). To achieve this objective, we conduct the now well- established experimental procedure, including (i) pre-training transformer-based GPT-2 LMs with Korean and English datasets, (ii) further fine-tuning them with a specific set of Korean data as L1 or L2, and (iii) evaluating them with the test data of KBLiMP while analyzing their linguistic generalization in L1 or L2. We have found negative transfer effects in the comparison between English as L1 and Korean as L2. Furthermore, in the trajectory analysis, the second language-learning LM has captured linguistic features of Korean including syntax, syntax-semantics interface, and morphology during the progressive training step. Our study of second language learning in LMs contributes to predicting potential syntactic challenges arising from the interference by the L1 language during the learning of Korean as a foreign language.

L1 Influence on the Use of the English Present Perfect: A Corpus Analysis of Russian and Spanish Learners’ Essays

Background: Mastering verbal tenses, especially those expressing aspect, in a second language presents a challenge as learners frequently link the semantic nuances of verbal forms in their second language (L2) to the characteristics of the verbal systems in their native languages (L1). This study explores the impact of L1 on the usage of the English Present Perfect (PP) among non-native speakers. Purpose: In an effort to contribute to the ongoing research on the mechanisms governing the acquisition of English tenses, this study focuses on the variations that affect the usage of the PP in the writing of English learners. The investigation is particularly centered on university students whose L1 is Russian and Spanish, seeking to delve into the ways in which their first language influences the utilisation of the PP in their English writing. Method: Analysis of L2 English by Russian and Spanish learners, based on corpora of argumentative essays written by undergraduate Russian and Spanish learners of English, controlled by a corpus of essays produced by native speakers of English; frequency and distribution of the PP in learner writings; examination of semantic contexts; identification of error types. Results: The findings indicate that, despite a higher occurrence of the PP in texts produced by Spanish learners compared to Russian learners, the rate of errors in its application is nearly identical in both learner corpora. These errors are likely attributable to challenges in comprehending the functions of the PP and in distinguishing its semantics from those of other English tenses, particularly the Past Simple. Conclusion: The study suggests that the increased prevalence of PP usage by L2 learners may be attributed to positive transfer from their L1 when it exhibits structures analogous to the English PP. Conversely, patterns indicative of, for example, undergeneralisation of semantic contexts suggesting the relevance of an action, or of overgeneralisation of adverbs compatible with the PP can be interpreted as evidence of negative transfer. The results of this study hold significance for language pedagogy, as they highlight potential challenges in acquiring the PP that learners from diverse L1 backgrounds may encounter.

A novel population robustness-based switching response framework for solving dynamic multi-objective problems

In this paper, a novel population robustness-based switching response framework (PR-SRF) is proposed to develop effective dynamic multi-objective optimization algorithm (DMOA), which integrates different response strategies to comprehensively cope with the dynamic behaviors. In particular, the population robustness is described by the quantification of how severely the environmental changes affect current population, which is timely graded as three levels of weak, strong, and normal to enable the adaptive switch of three different responses of diversity-enhancement, diversity-maintenance, and the knowledge-transfer, respectively. In this way, associations between the adopted responses and the changing environments are successfully established, thereby facilitating more intelligent decision when handling the dynamic behaviors. According to the benchmark evaluation results, the proposed PR-SRF-DMOA yields better comprehensive performance than several other DMOAs with popular response strategies, and it also outperforms another three DMOAs with hybrid responses, which demonstrates the great competitiveness of our algorithm. In addition, ablation study proves that the proposed PR-SRF can sufficiently exploit the merits of different responses, which effectively alleviates the negative knowledge transfer in extremely fluctuating environments, thereby providing valuable references for the development of evolutionary transfer optimization (ETO) algorithms.

A multi-fidelity transfer learning strategy based on multi-channel fusion

Multi-fidelity strategies leverage a large amount of low-fidelity data combined with a smaller set of high-fidelity data, thereby achieving satisfactory results at a reasonable cost. In our research, we introduce an innovative multi-fidelity strategy that integrates the concepts of multi-fidelity data fusion and transfer learning. In the proposed framework, we incorporate auto-encoders and a multi-channel transfer learning strategy, enabling the network model to comprehend the relationship between the low-fidelity and high-fidelity models in both explicit and implicit manners. This approach not only enhances prediction accuracy but also mitigates issues such as overfitting and negative transfer, which may arise in scenarios with sparse samples. Additionally, Bayesian optimization is employed for effective hyperparameter selection. To evaluate and analyze the performance of our proposed method, we present a series of benchmark test cases. Furthermore, we also show the application of the proposed method to engineering problems. Firstly, we consider a parametrized partial differential equation problem, where high-fidelity and low-fidelity data are obtained using exact methods and simplified algorithms, respectively. Subsequently, we extend this strategy to convolutional neural network architectures, specifically addressing a pressure Poisson equation problem. We also explore the effect of the reliability of the low-fidelity data and the number of high-fidelity data on the results. The results show that the proposed method exhibits low requirements in terms of both the reliability of the low-fidelity data and the number of high-fidelity data while maintaining satisfactory accuracy metrics.

Open Set Bearing Fault Diagnosis with Domain Adaptive Adversarial Network under Varying Conditions

Bearing fault diagnosis is a pivotal aspect of monitoring rotating machinery. Recently, numerous deep learning models have been developed for intelligent bearing fault diagnosis. However, these models have typically been established based on two key assumptions: (1) that identical fault categories exist in both the training and testing datasets, and (2) the datasets used for testing and training are assumed to follow the same distribution. Nevertheless, these assumptions prove impractical and fail to accurately depict real-world scenarios, particularly those involving open-world assumption fault diagnosis in multi-condition scenarios. For that purpose, an open set domain adaptive adversarial network framework is proposed. Specifically, in order to improve the learning of distribution characteristics in different fields, comprehensive training is implemented using a deep convolutional autoencoder model. Additionally, to mitigate the negative transfer resulting from unknown fault samples in the target domain, the similarity of each target domain sample and the shared classes in the source domain are estimated using known class classifiers and extended classifiers. Similarity weight values are assigned to each target domain sample, and an unknown boundary is established in a weighted manner. This approach is employed to establish the alignment between the classes shared between the two domains, enabling the classification of known fault classes, while allowing the recognition of unknown fault classes in the target domain. The efficacy of our suggested approach is empirically validated using different datasets.

STEM: Unleashing the Power of Embeddings for Multi-Task Recommendation

Multi-task learning (MTL) has gained significant popularity in recommender systems as it enables simultaneous optimization of multiple objectives. A key challenge in MTL is negative transfer, but existing studies explored negative transfer on all samples, overlooking the inherent complexities within them. We split the samples according to the relative amount of positive feedback among tasks. Surprisingly, negative transfer still occurs in existing MTL methods on samples that receive comparable feedback across tasks. Existing work commonly employs a shared-embedding paradigm, limiting the ability of modeling diverse user preferences on different tasks. In this paper, we introduce a novel Shared and Task-specific EMbeddings (STEM) paradigm that aims to incorporate both shared and task-specific embeddings to effectively capture task-specific user preferences. Under this paradigm, we propose a simple model STEM-Net, which is equipped with an All Forward Task-specific Backward gating network to facilitate the learning of task-specific embeddings and direct knowledge transfer across tasks. Remarkably, STEM-Net demonstrates exceptional performance on comparable samples, achieving positive transfer. Comprehensive evaluation on three public MTL recommendation datasets demonstrates that STEM-Net outperforms state-of-the-art models by a substantial margin. Our code is released at https://github.com/LiangcaiSu/STEM.

On Disentanglement of Asymmetrical Knowledge Transfer for Modality-Task Agnostic Federated Learning

There has been growing concern regarding data privacy during the development and deployment of Multimodal Foundation Models for Artificial General Intelligence (AGI), while Federated Learning (FL) allows multiple clients to collaboratively train models in a privacy-preserving manner. This paper formulates and studies Modality-task Agnostic Federated Learning (AFL) to pave the way toward privacy-preserving AGI. A unique property of AFL is the asymmetrical knowledge relationships among clients due to modality gaps, task gaps, and domain shifts between clients. This raises a challenge in learning an optimal inter-client information-sharing scheme that maximizes positive transfer and minimizes negative transfer for AFL. However, prior FL methods, mostly focusing on symmetrical knowledge transfer, tend to exhibit insufficient positive transfer and fail to fully avoid negative transfer during inter-client collaboration. To address this issue, we propose DisentAFL, which leverages a two-stage Knowledge Disentanglement and Gating mechanism to explicitly decompose the original asymmetrical inter-client information-sharing scheme into several independent symmetrical inter-client information-sharing schemes, each of which corresponds to certain semantic knowledge type learned from the local tasks. Experimental results demonstrate the superiority of our method on AFL than baselines.