Search

Simple Summary: Transfer learning is a technique utilizing a pre-trained model's knowledge in a new task. This helps reduce the sample size and time needed for training. These characteristics of transfer learning make it a perfect candidate to use in genetic research. The aim of our study is to review the current uses of transfer learning in genetic research. Here, we overview the use of transfer learning in the mutation detection of different cancers (lung, gastrointestinal, breast, glioma), gene expression, genetic syndrome detection (Down's syndrome, Noonan syndrome, Williams–Beuren syndrome) based on the phenotype of patients, and identifying possible genotype–phenotype association. Using transfer learning in model development increases the final performance of the model compared with models trained from scratch. Artificial intelligence (AI), encompassing machine learning (ML) and deep learning (DL), has revolutionized medical research, facilitating advancements in drug discovery and cancer diagnosis. ML identifies patterns in data, while DL employs neural networks for intricate processing. Predictive modeling challenges, such as data labeling, are addressed by transfer learning (TL), leveraging pre-existing models for faster training. TL shows potential in genetic research, improving tasks like gene expression analysis, mutation detection, genetic syndrome recognition, and genotype–phenotype association. This review explores the role of TL in overcoming challenges in mutation detection, genetic syndrome detection, gene expression, or phenotype–genotype association. TL has shown effectiveness in various aspects of genetic research. TL enhances the accuracy and efficiency of mutation detection, aiding in the identification of genetic abnormalities. TL can improve the diagnostic accuracy of syndrome-related genetic patterns. Moreover, TL plays a crucial role in gene expression analysis in order to accurately predict gene expression levels and their interactions. Additionally, TL enhances phenotype–genotype association studies by leveraging pre-trained models. In conclusion, TL enhances AI efficiency by improving mutation prediction, gene expression analysis, and genetic syndrome detection. Future studies should focus on increasing domain similarities, expanding databases, and incorporating clinical data for better predictions. [ABSTRACT FROM AUTHOR]

Previous research indicated a need to improve pilot training with regard to understanding of autopilot logic and behavior, especially in non-routine situations. Therefore, we tested the effect of problem-based exploratory training on pilots' understanding of autopilot functions. Using a moving-base flight simulator, general aviation pilots (n = 45) were trained to diagnose failures either without foreknowledge and guidance (exploratory group), without foreknowledge but with some guidance (exploratory-guidance group) or with foreknowledge and full guidance (control group). They subsequently performed six test scenarios in which their understanding of the effects of failures was tested by requiring them to deduce the failures and select autopilot modes that were still functioning. Those who received exploratory training with guidance were significantly more likely than the other groups to diagnose failures correctly. The exploratory training group also selected the most appropriate functioning autopilot modes significantly faster than the control group. The results suggest that exploratory training with an appropriate level of guidance is useful for gaining a practical understanding of autopilot logic and behavior. Exploratory training may help to improve transfer of training to operational practice, and prevent automation surprises and accidents. [ABSTRACT FROM AUTHOR]

Background: Prospective observational studies have demonstrated that the machine learning (ML) -guided noninvasive chromosome screening (NICS) grading system, which we called the noninvasive chromosome screening-artificial intelligence (NICS-AI) grading system, can be used embryo selection. The current prospective interventional clinical study was conducted to investigate whether this NICS-AI grading system can be used as a powerful tool for embryo selection. Methods: Patients who visited our centre between October 2018 and December 2021 were recruited. Grade A and B embryos with a high probability of euploidy were transferred in the NICS group. The patients in the control group selected the embryos according to the traditional morphological grading. Finally, 90 patients in the NICS group and 161 patients in the control group were compared statistically for their clinical outcomes. Results: In the NICS group, the clinical pregnancy rate (70.0% vs. 54.0%, p < 0.001), the ongoing pregnancy rate (58.9% vs. 44.7%, p = 0.001), and the live birth rate (56.7% vs. 42.9%, p = 0.001) were significantly higher than those of the control group. When the female was ≥ 35 years old, the clinical pregnancy rate (67.7% vs. 32.1%, p < 0.001), ongoing pregnancy rate (56.5% vs. 25.0%, p = 0.001), and live birth rate (54.8% vs. 25.0%, p = 0.001) in the NICS group were significantly higher than those of the control group. Regardless of whether the patients had a previous record of early spontaneous abortion or not, the live birth rate of the NICS group was higher than that of the control group (61.0% vs. 46.9%; 57.9% vs. 34.8%; 33.3% vs. 0%) but the differences were not statistically significant. Conclusions: NICS-AI was able to improve embryo utilisation rate, and the live birth rate, especially for those ≥ 35 years old, with transfer of Grade A embryos being preferred, followed by Grade B embryos. NICS-AI can be used as an effective tool for embryo selection in the future. [ABSTRACT FROM AUTHOR]

On-shelf availability (OSA) refers to the number of products available in saleable condition to a customer at the place he expects and at the time he wants to buy the product. OSA of the products in retail stores is important to enhance the customer shopping experience and profitability. The products on the retail store shelves are arranged strategically. This arrangement may be changed during shopping, because it is observed that customers may not move the products in the same or at the right place when they pick them up from the shelf. In such real-time scenarios, it would be appropriate to perform shelf image analysis to detect and identify these misplaced objects. This will help the shopkeeper to maintain a large-scale store effortlessly, otherwise it will require massive human effort and labour along with time. Many computer vision-based technology solutions have been developed by researchers to solve this problem. This article discusses a convolutional neural network (CNN)-based method for classifying shelf imagery into the correct semantic classes (that is whether it has misplaced products or not). The proposed architecture was evaluated with a modified COIL-100 dataset. The results of the proposed CNN model were compared with the variant of the proposed model in addition to lightweight pre-trained deep learning models viz; mobilenetV2, densenet121, efficientnetb0, efficientnetb3, and NASNetMobile using transfer learning (TL) concept. According to this study, it has been found that the TL-based MobilenetV2 model is lightweight and achieves better classification performance with 91.28% accuracy. Furthermore, a CNN-based model with 11 user-defined layers provides an accuracy of 90.36%. This shows the efficacy of the proposed models for change detection. [ABSTRACT FROM AUTHOR]

Deep subsurface exploration is important for mining, oil and gas industries, as well as in the assessment of geological units for the disposal of chemical or nuclear waste, or the viability of geothermal energy systems. Typically, detailed examinations of subsurface formations or units are performed on cuttings or core materials extracted during drilling campaigns, as well as on geophysical borehole data, which provide detailed information about the petrophysical properties of the rocks. Depending on the volume of rock samples and the analytical program, the laboratory analysis and diagnostics can be very time-consuming. This study investigates the potential of utilizing machine learning, specifically convolutional neural networks (CNN), to assess the lithology and mineral content solely from analysis of drill core images, aiming to support and expedite the subsurface geological exploration. The paper outlines a comprehensive methodology, encompassing data preprocessing, machine learning methods, and transfer learning techniques. The outcome reveals a remarkable 96.7% accuracy in the classification of drill core segments into distinct formation classes. Furthermore, a CNN model was trained for the evaluation of mineral content using a learning data set from multidimensional log analysis data (silicate, total clay, carbonate). When benchmarked against laboratory XRD measurements on samples from the cores, both the advanced multidimensional log analysis model and the neural network approach developed here provide equally good performance. This work demonstrates that deep learning and particularly transfer learning can support extracting petrophysical properties, including mineral content and formation classification, from drill core images, thus offering a road map for enhancing model performance and data set quality in image-based analysis of drill cores. [ABSTRACT FROM AUTHOR]

Aging is often accompanied by a decline in cognitive functions, with memory being particularly affected. Recent studies suggest that cognitive training sessions that teach memory strategies relevant to daily life may benefit seniors who live in the community. However, it is possible that the cognitive improvement observed in these programs results from the social encounters embedded in them. In this study, we aimed to investigate the effect of a social cognitive training group, which met regularly for an extended period, on enhancing cognitive indices compared to a control group that only received social engagement meetings without training. Sixty-six participants with a mean age of 78 took part in 12 sessions of a social engagement group, with or without strategy training. Cognitive performance was assessed before and after training using four memory tasks, two similar to the trained tasks (near-transfer tasks) and two novel (far-transfer tasks). Both groups showed a slight improvement in most of the evaluation tasks, but the cognitive training combined with social engagement group showed a significant improvement in the Word Recall and Verbal Fluency tests compared to the social engagement group without training. Our findings suggest that cognitive training sessions may be a useful tool in promoting cognitive improvement among older adults living in the community, even beyond the improvement obtained from the social engagement that occurs during the training sessions. Trial registration number: NCT05016336. Date of registration: 20 August 2021. Retrospectively registered. [ABSTRACT FROM AUTHOR]

Unsupervised cross-modality domain adaptation for medical image segmentation has made great progress with the development of adversarial learning-based methods, but the training of adversarial models is considerably complicated. In this paper, we propose a conceptually simple but effective unsupervised domain adaptation method to achieve adaptation on frequency spectrum components of target and source images decomposed by a novel spectrum transfer strategy. Specifically, we replace the high-frequency components of the source domain images with that of the target domain images for details feature adaptation and adjust the low-frequency components by histogram matching for style adaptation. Besides, we propose multi-direction collaborative learning on both target and source domains to further improve the performance. Experimental results demonstrate that our method significantly outperforms state-of-the-art UDA methods for medical image segmentation on two publicly available datasets (cardiac dataset, and abdominal multi-organ dataset) in both CT to MRI and MRI to CT domain adaptation scenarios [ABSTRACT FROM AUTHOR]

Background: Transfer students are a growing population of undergraduate psychology majors who are academically at-risk but often underserved by 4-year institutions. Objective: We describe a transfer learning community (TLC) for entering psychology transfer students and report its impact on participants social and academic integration and success. Method: A mixed-method design was used to compare academic success metrics (archival measures of first-semester grade point average and probation rates) of participants with entering transfer majors not participating in the learning community. A participant survey was also created and used to measure student perceptions of the TLC, as well as perceived challenges of transferring during the Covid-19 pandemic and how the TLC helped to mitigate them. Results: During its first year, the TLC served 72 entering transfer majors across two cohorts. Participants significantly out-performed non-participant psychology transfer majors on first-semester grade point average and were significantly less likely to be placed on academic probation after their first semester than non-participant transfer majors. Conclusion: The widely-reported benefits of learning communities for first-year psychology students are adaptable to the more at-risk population of transfer students. Teaching Implications: Technology-enabled teaching tools can be used to build virtual community in hybrid and synchronous online learning environments. [ABSTRACT FROM AUTHOR]

The elevator door system plays a crucial role in ensuring elevator safety. Fault prediction is an invaluable tool for accident prevention. By analyzing the sound signals generated during operation, such as component wear and tear, the fault of the system can be accurately determined. This study proposes a GNN-LSTM-BDANN deep learning model to account for variations in elevator operating environments and sound signal acquisition methods. The proposed model utilizes the historical sound data from other elevators to predict the remaining useful life (RUL) of the target elevator door system. Firstly, the opening and closing sounds of other elevators is collected, followed by the extraction of relevant sound signal characteristics including A-weighted sound pressure level, loudness, sharpness, and roughness. These features are then transformed into graph data with geometric structure representation. Subsequently, the Graph Neural Networks (GNN) and long short-term memory networks (LSTM) are employed to extract deeper features from the data. Finally, transfer learning based on the improved Bhattacharyya Distance domain adversarial neural network (BDANN) is utilized to transfer knowledge learned from historical sound data of other elevators to predict RUL for the target elevator door system effectively. Experimental results demonstrate that the proposed method can successfully predict potential failure timeframes for different elevator door systems. [ABSTRACT FROM AUTHOR]

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. [ABSTRACT FROM AUTHOR]

Despite significant progress has been made towards crop yield prediction with remote sensing, there exist knowledge gaps on (1) the impacts of temporal resolution of imaging frequencies on yield prediction, (2) transferability of the models among different genotypes and test sites, and (3) translation of these research developments to crop breeding that benefit farmers. Existing research predominantly provides an on-site perspective, frequently missing the complexities of real-world applications. The objectives of this paper are to investigate the transferability and generalization capabilities of yield prediction models for crop breeding across test sites located in North and South Americas. Toward that goal, we tested different machine learning techniques including Random Forest Regressor (RF), Gradient Boosting Regression (GB), and Deep Neural Networks (DNN) for soybean yield prediction with experiments conducted in different climate and growth conditions. A novel transfer learning approach was proposed for genotype selection and categorizing soybean yield for screening high-yield varieties. Furthermore, we studied the effect of temporal resolution on yield prediction, focusing on the critical development stages and optimal aerial survey frequencies for precise yield prediction using large 31,404 sample data. Results demonstrated that the combined dataset of Argentina and United States representing different climate regimes provided the highest performance with an R2 of 0.76 using RF and GB algorithms. The classification approach was proven to be most useful for crop breeding as demonstrated by accurately identifying the high-yielding genotypes. Increasing temporal sampling of key phenological stages significantly improved yield prediction. Although transfer learning yielded promising outcomes across trials within Argentina the efficacy of transferring models from Argentina to the United States was limited, attributed to significant seasonal and climate variations. This study pioneered the use of transfer learning for model adaptability in real-world breeding scenarios, training and transferring models within the South and North Americas, providing actionable insights and strategies for the breeding community, aiming to facilitate improved decision-making for agricultural productivity. [ABSTRACT FROM AUTHOR]

Considering the advantages of brain stimulation techniques in detecting the role of different areas of the brain in human sensorimotor behaviors, we used anodal transcranial direct-current stimulation (a-tDCS) over three different brain sites of the frontoparietal cortex (FPC) in healthy participants to elucidate the role of these three brain areas of the FPC on reaction time (RT) during a sequential visual isometric pinch task (SVIPT). We also aimed to assess if the stimulation of these cortical sites affects the transfer of learning during SVIPT. A total of 48 right-handed healthy participants were randomly assigned to one of the four a-tDCS groups: (1) left primary motor cortex (M1), (2) left dorsolateral prefrontal cortex (DLPFC), (3) left posterior parietal cortex (PPC), and (4) sham. A-tDCS (0.3 mA, 20 min) was applied concurrently with the SVIPT, in which the participants precisely controlled their forces to reach seven different target forces from 10 to 40% of the maximum voluntary contraction (MVC) presented on a computer screen with the right dominant hand. Four test blocks were randomly performed at the baseline and 15 min after the intervention, including sequence and random blocks with either hand. Our results showed significant elongations in the ratio of RTs between the M1 and sham groups in the sequence blocks of both the right-trained and left-untrained hands. No significant differences were found between the DLPFC and sham groups and the PPC and sham groups in RT measurements within the SVIPT. Our findings suggest that RT improvement within implicit learning of an SVIPT is not mediated by single-session a-tDCS over M1, DLPFC, or PPC. Further research is needed to understand the optimal characteristics of tDCS and stimulation sites to modulate reaction time in a precision control task such as an SVIPT. [ABSTRACT FROM AUTHOR]

The acquisition of skills, such as learning to play a musical instrument, involves various phases that make specific demands on the learner. Knowledge of the cognitive and motor contributions during learning phases can be helpful in developing effective and targeted interventions for healthy aging. Eighty-six healthy older participants underwent an extensive cognitive, motoric, and musical test battery. Within one session, one piano-related and one music-independent movement sequence were both learned. We tested the associations between skill performance and cognito-motor abilities with Bayesian mixed models accounting for individual learning rates. Results showed that performance was positively associated with all cognito-motor abilities. Learning a piano-related task was characterized by relatively strong initial associations between performance and abilities. These associations then weakened considerably before increasing exponentially from the second trial onwards, approaching a plateau. Similar performance–ability relationships were detected in the course of learning a music-unrelated motor task. Positive performance–ability associations emphasize the potential of learning new skills to produce positive cognitive and motor transfer effects. Consistent high-performance tasks that demand maximum effort from the participants could be very effective. However, interventions should be sufficiently long so that the transfer potential can be fully exploited. [ABSTRACT FROM AUTHOR]

Background: Depression often leads to clinical admission. Stimulating positive memory bias through repeated retrieval can counteract a ruminative depressotypic processing style. Cognitive Bias Modification (CBM)-Memory is a psychological intervention, which was evaluated as possible adjunct treatment for depressed inpatients. Methods: A randomized controlled pilot study with a one-week follow-up was implemented to compare the effect of a four-session positive CBM-Memory intervention to a neutral CBM-Memory version. Training-congruent retrieval (as manipulation check), and transfer to rumination (as mechanistic target) and depressive symptoms (self-rated and clinician-rated; indication of clinical relevance) were assessed. Results: In the intention-to-treat sample (N = 81, Mage=35.6/SD = 11.9, 60% female), condition-congruent memory retrieval of previously presented target words was found after the training and at follow-up (large effect-sizes). Positive CBM-Memory intervention resulted in sustained recall bias and largest rumination reduction with a small-medium effect size. Additionally, the conditions did not differ significantly on the reduction in depressive symptoms. Conclusions: Positive CBM-Memory intervention yielded positive recall bias, which was still present one week later. Additionally, the positive condition did yield more decrease in trait (but not state) rumination, providing some support for rumination as mechanistic target of positive CBM-Memory. However, there was no evidence for transfer of the training effect to depressive symptoms, indicating no superiority on clinical recovery of the positive over the neutral condition. Future research should examine a higher dosage, integration of CBM in the treatment provision, and its long-term effects in a well-powered trial. [ABSTRACT FROM AUTHOR]

Random noise suppression is an essential task in the seismic data processing. In recent years deep learning methods have achieved superior results in seismic data denoising. However, obtaining clean data from field seismic data for training is challenging. Therefore, supervised deep learning denoising methods can only use synthetic datasets or field datasets constructed by conventional seismic denoising methods for training. Aiming at this problem, we proposed a self-supervised deep learning seismic denoising method based on Neighbor2Neighbor. This method only requires sampling the noisy data twice to train the denoising network without clean data. For the characteristics of seismic data, we designed a vertical neighbor subsample to make Neighbor2Neighbor more suitable for seismic data. In addition, to further improve the denoising effect in field data, we introduced a transfer learning strategy in our method. Numerical experiments demonstrated that our method outperformed both the conventional denoising seismic method and the supervised learning seismic denoising method after transfer learning. [ABSTRACT FROM AUTHOR]

System knowledge and reasoning mechanisms are essential means for intellectualization of cyber-physical systems (CPSs). As enablers of system intelligence, they make such systems able to solve application problems and to maintain their efficient operation. Normally, system intelligence has a human-created initial part and a system-produced (extending) part, called synthetic system intelligence (SSI). This position paper claims that SSI can be converted to a new industrial asset and utilized as such. Unfortunately, no overall theory of SSI exists and its conceptual framework, management strategy, and computational methodologies are still in a premature stage. This is the main reason why no significant progress has been achieved in this field, contrary to the latent potentials. This paper intends to contribute to: (i) understanding the nature and fundamentals of SSI, (ii) systematizing the elicitation and transfer of SSI, (iii) exploration of analogical approaches to utilization of SSI, and (iv) road-mapping and scenario development for the exploitation of SSI as an industrial asset. First, the state of the art is surveyed and the major findings are presented. Then, four families of analogical approaches to SSI transfer are analyzed. These are: (i) knowledge transfer based on repositories, (ii) transfer among agents, (iii) transfer of learning resources, and (iv) transfer by emerging approaches. A procedural framework is proposed that identifies the generic functionalities needed for a quasi-autonomous handling of SSI as an industrial asset. The last section casts light on some important open issues and necessary follow-up research and development activities. [ABSTRACT FROM AUTHOR]

Assigning repetitive and physically demanding construction tasks to robots can alleviate human workers' exposure to occupational injuries, which often result in significant downtime or premature retirement. However, the successful delegation of construction tasks and the achievement of high-quality robot-constructed work requires transferring necessary dexterous and adaptive construction craft skills from workers to robots. Predefined motion planning scripts tend to generate rigid and collision-prone robotic behaviors in unstructured construction site environments. In contrast, imitation learning (IL) offers a more robust and flexible skill transfer scheme. However, the majority of IL algorithms rely on human workers repeatedly demonstrating task performance at full scale, which can be counterproductive and infeasible in the case of construction work. To address this concern, in this paper, we propose an immersive and Cloud Robotics-based virtual demonstration framework that serves two primary purposes. First, it digitalizes the demonstration process, eliminating the need for repetitive physical manipulation of heavy construction objects. Second, it employs a federated collection of reusable demonstrations that are transferable for similar tasks in the future and can, consequently, reduce the requirement for repetitive illustration of tasks by human agents. In addition, to enhance the trustworthiness, explainability, and ethical soundness of the robot training, this framework utilizes a hierarchical imitation learning (HIL) model to decompose human manipulation skills into sequential and reactive subskills. These two layers of skills are represented by deep generative models; these models enable adaptive control of robot action. The proposed framework has the potential to mitigate technical adoption barriers and facilitate the practical deployment of full-scale construction robots to perform a variety of tasks with human supervision. By delegating the physical strains of construction work to human-trained robots, this framework promotes the inclusion of workers with diverse physical capabilities and educational backgrounds within the construction industry. [ABSTRACT FROM AUTHOR]

Artificial intelligence-powered deep learning methods are being used to diagnose brain tumors with high accuracy, owing to their ability to process large amounts of data. Magnetic resonance imaging stands as the gold standard for brain tumor diagnosis using machine vision, surpassing computed tomography, ultrasound, and X-ray imaging in its effectiveness. Despite this, brain tumor diagnosis remains a challenging endeavour due to the intricate structure of the brain. This study delves into the potential of deep transfer learning architectures to elevate the accuracy of brain tumor diagnosis. Transfer learning is a machine learning technique that allows us to repurpose pre-trained models on new tasks. This can be particularly useful for medical imaging tasks, where labelled data is often scarce. Four distinct transfer learning architectures were assessed in this study: ResNet152, VGG19, DenseNet169, and MobileNetv3. The models were trained and validated on a dataset from benchmark database: Kaggle. Five-fold cross validation was adopted for training and testing. To enhance the balance of the dataset and improve the performance of the models, image enhancement techniques were applied to the data for the four categories: pituitary, normal, meningioma, and glioma. MobileNetv3 achieved the highest accuracy of 99.75%, significantly outperforming other existing methods. This demonstrates the potential of deep transfer learning architectures to revolutionize the field of brain tumor diagnosis. [ABSTRACT FROM AUTHOR]

Introduction: Perception under noisy conditions requires not only feature identification but also a process whereby target features are selected and noise is filtered out (e.g., when identifying an animal hiding in the savannah). Interestingly, previous perceptual learning studies demonstrated the utility of training feature representation (without noise) for improving discrimination under noisy conditions. Furthermore, learning to filter out noise also appears to transfer to other perceptual task under similar noisy conditions. However, such learning transfer effects were thus far demonstrated predominantly in simple stimuli. Here we sought to explore whether similar learning transfer can be observed with complex real-world stimuli. Methods: We assessed the feature-to-noise transfer effect by using complex stimuli of human faces. We first examined participants' performance on a face-noise task following either training in the same task, or in a different face-feature task. Second, we assessed the transfer effect across different noise tasks defined by stimulus complexity, simple stimuli (Gabor) and complex stimuli (faces). Results: We found a clear learning transfer effect in the face-noise task following learning of face features. In contrast, we did not find transfer effect across the different noise tasks (from Gabor-noise to face-noise). Conclusion: These results extend previous findings regarding transfer of feature learning to noisy conditions using real-life stimuli. [ABSTRACT FROM AUTHOR]

This paper presents an innovative approach to the automatic detection of rice leaf diseases using advanced deep learning techniques, specifically focusing on the use of convolutional neural networks (CNNs) augmented with the power of transfer learning. The research compiles a comprehensive dataset of rice leaf images depicting several disease states such as leaf blast, leaf blight, brown spot, as well as healthy leaf conditions. Each disease exhibits distinct visual symptoms: leaf blast is identified by its elongated lesions with darkened perimeters, leaf blight by its irregularly edged lesions, and brown spot by its circular lesions encircled by a brownish hue. The challenge of analyzing images with varying levels of clarity and contrast, particularly in healthy leaf samples, is addressed through the application of transfer learning. This enhances the precision of disease classification within our model. Utilizing well-known algorithms like MobileNet and VGG16, the proposed system efficiently differentiates among various rice leaf ailments, including Bacterial Leaf Blight, Brown Spot, Healthy, Leaf Blast, Leaf Scald, and Narrow Brown Spot. The results demonstrate the model's effectiveness in accurately detecting different rice leaf diseases, highlighting its potential to improve disease management strategies in agriculture and increase crop production efficiency. [ABSTRACT FROM AUTHOR]

Quantitative analysis of rock samples using laser-induced breakdown spectroscopy (LIBS) is a challenging task due to the significant differences in spectra between pressed pellet samples and rock samples, with pressed pellet samples usually exhibiting stronger spectral lines. The predictive ability of LIBS quantitative analysis models for rock samples is poorer compared to that for pressed pellet samples, and quantitative analysis models constructed using pressed pellet samples cannot be directly applied to predict the composition of rock samples. To address the issue, a quantitative analysis method for raw rocks based on feature transfer learning using transfer component analysis (TCA) is proposed in this paper. By establishing a feature mapping relationship between pressed pellet samples and rock samples, the differences between data features are reduced, thus enabling the training of a more quantitative analysis model. The transfer learning was introduced into a multivariate regression machine learning model for training using pressed pellet samples, which successfully addresses the problem of prediction accuracy of rock sample element contents. In this model, all data were mapped to a high-dimensional reproducing kernel Hilbert space, and a subset of the most similar features was selected to train the quantitative analysis model. Upon testing the model with independent rock samples, the difference between the predicted and true values was significantly reduced. The model yielded a root mean square error of prediction (RMSEP) of 3.7131, 1.0185, 0.2985, 13.0439, and 1.5450 for Si, Al, Fe, Ca, and Mg in rock samples, respectively. These results indicate that LIBS coupled with the transfer learning algorithm can effectively eliminate the differences between the spectra of the pressed pellet and the raw rock and provide another idea for in situ LIBS detection. [ABSTRACT FROM AUTHOR]

Alzheimer's disease (AD) is the most frequent type of dementia that has no effective cure, except early discovery and treatment that may help patients to include successful years in patient's lives. Currently, mini-mental state examination (MMSE) score and manual examination of magnetic resource imaging (MRI) scan along with machine learning techniques are used to diagnose the disease; however, they possess certain accuracy limits. Therefore, this paper proposes a deep learning-based multilayered framework for AD classification using transfer learned Alexnet and LSTM for multiclass and binary classification of MR images. However, the deep learning models used in the current study necessitate a large training dataset to produce better outcomes. As a result, this work also utilizes generative adversarial network (GAN) as a data augmentation tool to improve the classification results and further to solve the problem of overfitting. The study uses Alzheimer's disease neuroimaging initiative (ADNI) dataset of 60 AD, 73 mild cognitive impairment (MCI) and 67 cognitively normal (CN) patients from which 2 D MR image scans are extracted. Furthermore, the proposed method achieved the classification accuracy on AD–CN at 98.13%, AD–MCI at 99.38% and CN–MCI at 99.37%, respectively. Also, the multiclass classification shows the promising accuracy of 96.83% for the proposed framework. Finally, the proposed model's performance is compared to other state-of-the-art techniques and the experimental results show that the proposed model outperforms in terms of accuracy, sensitivity and hypothesis testing. [ABSTRACT FROM AUTHOR]

The graph collaborative filtering algorithm, known for its excellent data modeling capability, has found wide applications in recommendation systems. However, traditional graph collaborative filtering algorithms often face challenges related to sparse interaction data and inefficient utilization of side information between users and items. To address these issues, a graph Contrastive Learning with Knowledge Transfer (CLKT) has been proposed. The proposed method aims to integrate the side information of user and item attributes with interaction data using GNN. By leveraging contrastive learning, it achieves cross-view learning within the graph collaborative filtering algorithm. Recognizing that the impact of side information on user-item interaction data can vary, the proposed method introduces feature transformers and knowledge transfer networks into contrastive learning. This enables the adaptive embedding of node features for users and items in CLKT. To further enhance the embedding capability for different node relationships, the CLKT method incorporates cluster-based contrastive learning. This approach considers the diverse impacts that different types of node relationships can have on node feature learning. By simultaneously learning both the interaction relationships between nodes and the clustering relationships among nodes, the proposed method enables the exploration of multiple-node relationships for feature mining. Extensive experiments on three real datasets demonstrate the effectiveness and feasibility of the proposed CLKT, and the results demonstrated its promising performance. [ABSTRACT FROM AUTHOR]

Although transfer models are limited in their ability to evolve over time and account for a wide range of processes, they have repeatedly shown to be useful for testing categorization theories and predicting participants' generalization performance. In this study, we propose a statistical framework that allows transfer models to be applied to category learning data. Our framework uses a segmentation/clustering technique specifically tailored to suit category learning data. We applied this technique to a well-known transfer model, the Generalized Context Model, in three novel experiments that manipulated ordinal effects in category learning. The difference in performance across the three contexts, as well as the benefit of the rule-based order observed in two out of three experiments, were mostly detected by the segmentation/clustering method. Furthermore, the analysis of the segmentation/clustering outputs using the backward learning curve revealed that participants' performance suddenly improved, suggesting the detection of an "eureka" moment. Our adjusted segmentation/clustering framework allows transfer models to fit learning data while capturing relevant patterns. [ABSTRACT FROM AUTHOR]

Executive Functions are a set of interrelated, top-down processes essential for adaptive goal-directed behaviour, frequently impaired across different neurodevelopmental disorders with variable degrees of severity. Many executive-function-training studies in children with neurodevelopmental disorders have focused on near effects, investigating post-treatment improvements on directly trained processes, while enhancements of skills not directly trained, defined as far effects, are less considered, albeit these could be extremely relevant for reducing the negative impact of a disorder's core symptomatology. This systematic review and metanalysis aims to investigate the far effect outcomes after EF training in children with different types of neurodevelopmental disorders. 17 studies met the inclusion criteria for the systematic review, while 15 studies were selected in the metanalysis. An overall statistically significant effect size was found in the majority of far effect outcome measures considered in the studies. In particular, trainings on executive functions determine significant far effects on daily life functioning (0.46, 95% CI: [0.05–0.87]) and clinical symptoms (0.33, 95% CI: [0.15–0.51]). Despite a high variability of the results, intensity, frequency and the laboratory/life contexts dimension seem to be the most influential variables in determining far effects. This systematic review and metanalysis highlights the need to measure far effects of executive function training in neurodevelopmental disorders, selecting treatments not only on directly targeted processes, but also according to far impacts on the functional weakness of the disorder. [ABSTRACT FROM AUTHOR]

Virtual environments allow teachers to approach the teaching process inclusively and creatively, to facilitate the effective transfer of learning and to support student learning with new technologies regardless of various barriers. The hereby study evaluates the students’ views concerning the virtual delivery of courses during their learning process taking into consideration their former face-to-face teaching-learning experience. It was based on the theoretical part which included typical aspects of online teaching and learning are presented and the most recent studies in this area are reviewed. Following previous research, we aim to develop an innovative model of successful online teaching and learning in higher education. A questionnaire-based research methodology was designed for 355 students at a university in the North-Eastern region of Romania. The questionnaire with closed-ended items was used for assessing students’ attitudes on the online teaching activity. The results indicate students’ preference for online courses, as well as a number of significant differences, depending on the study specialisation or form of education. Any future changes to online teaching and learning in universities should also take into account students’ and teachers’ views on this particular issue. Such an approach should allow us to see whether and how effective the use of virtual teaching environments can be for both students and teachers in higher education (HE). [ABSTRACT FROM AUTHOR]

Accurate brain tumor classification is crucial for enhancing the diagnosis, prognosis, and treatment of glioblastoma patients. We employed the ResNet101 deep learning method with transfer learning to analyze the 2021 Radiological Society of North America (RSNA) Brain Tumor challenge dataset. This dataset comprises four structural magnetic resonance imaging (MRI) sequences: fluid‐attenuated inversion‐recovery (FLAIR), T1‐weighted pre‐contrast (T1w), T1‐weighted post‐contrast (T1Gd), and T2‐weighted (T2). We assessed the model's performance using standard evaluation metrics. The highest performance to detect MGMT methylation status for patients suffering glioblastoma was an accuracy (85.48%), sensitivity (80.64%), specificity (90.32%). Whereas classification performance with no tumor was yielded with accuracy (85.48%), sensitivity (90.32%), specificity (80.64%). The radiomic features (74) computed with ensembled Bagged Tree and relief feature selection method (30/74) improved the validation accuracy of 84.3% and AUC of 0.9038 to detect. O6‐methylguanine‐DNA methyltransferase (MGMT) promoter methylation status in glioblastoma patients holds promise for optimizing treatment planning and prognosis. By understanding MGMT methylation status, clinicians can make informed decisions about treatment strategies, potentially leading to improved clinical outcomes. [ABSTRACT FROM AUTHOR]

This paper outlines a new approach to designing tunable electromagnetic (EM) graphene-based meta-surfaces using convolutional neural networks (CNNs). EM meta-surfaces have previously been used to manipulate EM waves by adjusting the local phase of subwavelength elements within the wavelength scale, resulting in a variety of intriguing devices. However, the majority of these devices have only been capable of performing a single function, making it difficult to achieve multiple functionalities in a single design. Graphene, as an active material, offers unique properties, such as tunability, making it an excellent candidate for achieving tunable meta-surfaces. The proposed procedure involves using two CNNs to design the passive structure of the graphene meta-surfaces and predict the chemical potentials required for tunable responses. The CNNs are trained using transfer learning, which significantly reduced the time required to collect the training dataset. The proposed inverse design methodology demonstrates excellent performance in designing reconfigurable EM meta-surfaces, which can be tuned to produce multiple functions, making it highly valuable for various applications. The results indicate that the proposed approach is efficient and accurate and provides a promising method for designing reconfigurable intelligent surfaces for future wireless communication systems. [ABSTRACT FROM AUTHOR]

When we intensively train a timing skill, such as learning to play the piano, we not only produce brain changes associated with task-specific learning but also improve our performance in other temporal behaviors that depend on these tuned neural resources. Since the neural basis of time learning and generalization is still unknown, we measured the changes in neural activity associated with the transfer of learning from perceptual to motor timing in a large sample of subjects (n = 65; 39 women). We found that intense training in an interval discrimination task increased the acuity of time perception in a group of subjects that also exhibited learning transfer, expressed as a reduction in inter-tap interval variability during an internally driven periodic motor task. In addition, we found subjects with no learning and/or generalization effects. Notably, functional imaging showed an increase in pre-supplementary motor area and caudate-putamen activity between the post- and pre-training sessions of the tapping task. This increase was specific to the subjects that generalized their timing acuity from the perceptual to the motor context. These results emphasize the central role of the cortico-basal ganglia circuit in the generalization of timing abilities between tasks. [ABSTRACT FROM AUTHOR]

Breast cancer remains a significant global public health concern, emphasizing the critical role of accurate histopathological analysis in diagnosis and treatment planning. In recent years, the advent of deep learning techniques has showcased notable potential in elevating the precision and efficiency of histopathological data analysis. The proposed work introduces a novel approach that harnesses the power of Transfer Learning to capitalize on knowledge gleaned from pre-trained models, adapting it to the nuanced landscape of breast cancer histopathology. Our proposed model, a Transfer Learning-based concatenated model, exhibits substantial performance enhancements compared to traditional methodologies. Leveraging well-established pretrained models such as VGG-16, MobileNetV2, ResNet50, and DenseNet121—each Convolutional Neural Network architecture designed for classification tasks—this study meticulously tunes hyperparameters to optimize model performance. The implementation of a concatenated classification model is systematically benchmarked against individual classifiers on histopathological data. Remarkably, our concatenated model achieves an impressive training accuracy of 98%. The outcomes of our experiments underscore the efficacy of this four-level concatenated model in advancing the accuracy of breast cancer histopathological data analysis. By synergizing the strengths of deep learning and transfer learning, our approach holds the potential to augment the diagnostic capabilities of pathologists, thereby contributing to more informed and personalized treatment planning for individuals diagnosed with breast cancer. This research heralds a promising stride toward leveraging cutting-edge technology to refine the understanding and management of breast cancer, marking a significant advancement in the intersection of artificial intelligence and healthcare. [ABSTRACT FROM AUTHOR]

We aimed to investigate transfer of learning, whereby previously acquired skills impact new task learning. While it has been debated whether such transfer may yield positive, negative, or no effects on performance, very little is known about the underlying neural mechanisms, especially concerning the role of inhibitory (GABA) and excitatory (Glu) (measured as Glu + glutamine (Glx)) neurometabolites, as measured by magnetic resonance spectroscopy (MRS). Participants practiced a bimanual coordination task across four days. The Experimental group trained a task variant with the right hand moving faster than the left (Task A) for three days and then switched to the opposite variant (Task B) on Day4. The control group trained Task B across four days. MRS data were collected before, during, and after task performance on Day4 in the somatosensory (S1) and visual (MT/V5) cortex. Results showed that both groups improved performance consistently across three days. On Day4, the Experimental group experienced performance decline due to negative task transfer while the control group continuously improved. GABA and Glx concentrations obtained during task performance showed no significant group-level changes. However, individual Glx levels during task performance correlated with better (less negative) transfer performance. These findings provide a first window into the neurochemical mechanisms underlying task transfer. [ABSTRACT FROM AUTHOR]

Reinforcement learning is an important technique in various fields, particularly in automated machine learning for reinforcement learning (AutoRL). The integration of transfer learning (TL) with AutoRL in combinatorial optimization is an area that requires further research. This paper employs both AutoRL and TL to effectively tackle combinatorial optimization challenges, specifically the asymmetric traveling salesman problem (ATSP) and the sequential ordering problem (SOP). A statistical analysis was conducted to assess the impact of TL on the aforementioned problems. Furthermore, the Auto_TL_RL algorithm was introduced as a novel contribution, combining the AutoRL and TL methodologies. Empirical findings strongly support the effectiveness of this integration, resulting in solutions that were significantly more efficient than conventional techniques, with an 85.7% improvement in the preliminary analysis results. Additionally, the computational time was reduced in 13 instances (i.e., in 92.8% of the simulated problems). The TL-integrated model outperformed the optimal benchmarks, demonstrating its superior convergence. The Auto_TL_RL algorithm design allows for smooth transitions between the ATSP and SOP domains. In a comprehensive evaluation, Auto_TL_RL significantly outperformed traditional methodologies in 78% of the instances analyzed. [ABSTRACT FROM AUTHOR]

Objective: The present scoping review aims to transform the diverse field of research on the effects of mixed reality-based training on performance in manual assembly tasks into comprehensive statements about industrial needs for and effects of mixed reality-based training. Background: Technologies such as augmented and virtual reality, referred to as mixed reality, are seen as promising media for training manual assembly tasks. Nevertheless, current literature shows partly contradictory results, which is due to the diversity of the hardware used, manual assembly tasks as well as methodological approaches to investigate the effects of mixed reality-based training. Method: Following the methodological approach of a scoping review, we selected 24 articles according to predefined criteria and analyzed them concerning five key aspects: (1) the needs in the industry for mixed reality-based training, (2) the actual use and classification of mixed reality technologies, (3) defined measures for evaluating the outcomes of mixed reality-based training, (4) findings on objectively measured performance and subjective evaluations, as well as (5) identified research gaps. Results: Regarding the improvement of performance and effectiveness through mixed reality-based training, promising results were found particularly for augmented reality-based training, while virtual reality-based training is mostly—but not consistently—as good as traditional training. Application: Mixed reality-based training is still not consistently better, but mostly at least as good as traditional training. However, depending on the use case and technology used, the training outcomes in terms of assembly performance and subjective evaluations show promising results of mixed reality-based training. [ABSTRACT FROM AUTHOR]

Vulnerabilities such as SQL injection represent a serious challenge to security. While tools with a pre-defined logic are commonly used in the field of penetration testing, the continually evolving nature of the security challenge calls for models able to learn autonomously from experience. In this paper we build on previous results on the development of reinforcement learning models devised to exploit specific forms of SQL injection, and we design agents that are able to tackle a varied range of SQL injection vulnerabilities, virtually comprising all the archetypes normally considered by experts. We show that our agents, trained on a synthetic environment, perform a transfer of learning among the different SQL injections challenges; in particular, they learn to use their queries to efficiently gain knowledge about multiple vulnerabilities at once. We also introduce a novel and more versatile way to interpret server messages that reduces reliance on expert inputs. Our simulations show the feasibility of our approach which easily deals with a number of homogeneous challenges, as well as some of its limitations when presented with problems having higher degrees of uncertainty. [ABSTRACT FROM AUTHOR]

Discrimination between tectonic earthquakes (EQs) and quarry blasts is important for accurate EQ cataloguing and seismic hazard analysis. However, reliable classification of these two types of seismic events is challenging with no prior knowledge of source parameters. Here, we applied deep learning to perform this classification task in southern California and eastern Kentucky. Since the two regions differ significantly in available labelled data, class imbalance and waveform characteristics, we adopted different strategies for them. We directly trained a convolutional neural network (CNN) for southern California due to its data abundancy. To alleviate the class imbalance, the blast data were augmented by repeated sampling. The model for California yields F1-scores of >83.5 per cent when estimated by individual stations and >98.1 per cent by network average (i.e. averaging the CNN's outputs on all available stations for each event). As eastern Kentucky has a much smaller data size, we apply transfer learning to the pre-trained California model to fit the Kentucky data. The transfer-learned model yields F1-scores of >86.9 per cent when estimated by individual stations and >96.7 per cent by network average. The transfer-learned model outperforms the model re-trained from scratch for the Kentucky data. Gradient-weighted class activation mapping shows the S onset and the S long-period coda are important to identify EQs and blasts, respectively. By visual inspections of the seismograms, the source locations, the origin time and the P -wave polarities, we verified that most of the events falsely predicted by our models are actually mislabelled by seismic analysts. Our models thus show great potential in helping seismic analysts find those mislabelled events which remain hidden in the historical catalogue. Our results demonstrate that deep learning can achieve high accuracy in seismic event discrimination and that transfer learning is effective and efficient to generalize deep learning models across different regions. [ABSTRACT FROM AUTHOR]

Objective: As people with HIV (PWH) age, they are at-risk of developing cognitive impairments compared to their seronegative counterparts. Although speed of processing (SOP) training may help improve this cognitive ability, less work has examined transfer to other cognitive domains. This study examined the effect of SOP training has on secondary cognitive domains in PWH aged 40+ years. Method: In this 3-group 2-year longitudinal study, 216 PWH with HIV-associated neurocognitive disorder (HAND) or borderline HAND were randomized to either: (1) 10 h of SOP training (n = 70); (2) 20 h of SOP training (n = 73); or (3) 10 h of an active control training (n = 73). Participants completed a comprehensive cognitive battery at baseline, immediately after training, and at 1 and 2 years. This battery yielded global and domain specific T-scores as well as a cognitive impairment variable. Generalized linear mixed-effect models were fitted to estimate between-group mean differences at the follow-up time-points adjusted for baseline. Results: No clinically or statistically significant improvements in any of the cognitive outcomes were observed. A sensitivity analysis was conducted; conclusions replicated those of the main analysis, with two exceptions: Global Function T and Psychomotor Speed T showed relevant training improvements among the intervention groups over the control group at the immediate post time point. Conclusions: Although SOP training has been shown to improve cognitive abilities that correspond to driving and mobility, such training has limited therapeutic utility in improving cognition in other domains in PWH with HAND. [ABSTRACT FROM AUTHOR]

This study presents an intelligent identification method for shale crack networks based on transfer learning. Focuses on investigating the main physical parameters of shale similar materials based on the similarity theory and the physical parameters of shale. Fracture network images obtained from shale-like materials. The fracture network images are then preprocessed using image processing technology to generate a high-quality shale image dataset. A deep learning transfer recognition model, based on ResNet-50, is constructed to detect model performance using these shale fracture network images. Experimental results and reliability analyses demonstrate that the ResNet-50 based deep learning transfer model achieves an accuracy of 87% for shale images, indicating high recognition accuracy and fast model convergence. The proposed intelligent shale crack identification method exhibits robustness and generalization ability, making it suitable for efficient fracture identification in geological, road, and bridge deck projects. [ABSTRACT FROM AUTHOR]

Inference of transfer operators from data is often formulated as a classical problem that hinges on the Ulam method. The conventional description, known as the Ulam–Galerkin method, involves projecting onto basis functions represented as characteristic functions supported over a fine grid of rectangles. From this perspective, the Ulam–Galerkin approach can be interpreted as density estimation using the histogram method. In this study, we recast the problem within the framework of statistical density estimation. This alternative perspective allows for an explicit and rigorous analysis of bias and variance, thereby facilitating a discussion on the mean square error. Through comprehensive examples utilizing the logistic map and a Markov map, we demonstrate the validity and effectiveness of this approach in estimating the eigenvectors of the Frobenius–Perron operator. We compare the performance of histogram density estimation (HDE) and kernel density estimation (KDE) methods and find that KDE generally outperforms HDE in terms of accuracy. However, it is important to note that KDE exhibits limitations around boundary points and jumps. Based on our research findings, we suggest the possibility of incorporating other density estimation methods into this field and propose future investigations into the application of KDE-based estimation for high-dimensional maps. These findings provide valuable insights for researchers and practitioners working on estimating the Frobenius–Perron operator and highlight the potential of density estimation techniques in this area of study. [ABSTRACT FROM AUTHOR]

Transferring knowledge from pre-trained deep models for downstream tasks, particularly with limited labeled samples, is a fundamental problem in computer vision research. Recent advances in large-scale, task-agnostic vision-language pre-trained models, which are learned with billions of samples, have shed new light on this problem. In this study, we investigate how to efficiently transfer aligned visual and textual knowledge for downstream visual recognition tasks. We first revisit the role of the linear classifier in the vanilla transfer learning framework, and then propose a new paradigm where the parameters of the classifier are initialized with semantic targets from the textual encoder and remain fixed during optimization. To provide a comparison, we also initialize the classifier with knowledge from various resources. In the empirical study, we demonstrate that our paradigm improves the performance and training speed of transfer learning tasks. With only minor modifications, our approach proves effective across 17 visual datasets that span three different data domains: image, video, and 3D point cloud. [ABSTRACT FROM AUTHOR]

Self-supervised learning (SSL) has emerged as a promising paradigm for learning flexible speech representations from unlabeled data. By designing pretext tasks that exploit statistical regularities, SSL models can capture useful representations that are transferable to downstream tasks. Barlow Twins (BTs) is an SSL technique inspired by theories of redundancy reduction in human perception. In downstream tasks, BTs representations accelerate learning and transfer this learning across applications. This study applies BTs to speech data and evaluates the obtained representations on several downstream tasks, showing the applicability of the approach. However, limitations exist in disentangling key explanatory factors, with redundancy reduction and invariance alone being insufficient for factorization of learned latents into modular, compact, and informative codes. Our ablation study isolated gains from invariance constraints, but the gains were context-dependent. Overall, this work substantiates the potential of Barlow Twins for sample-efficient speech encoding. However, challenges remain in achieving fully hierarchical representations. The analysis methodology and insights presented in this paper pave a path for extensions incorporating further inductive priors and perceptual principles to further enhance the BTs self-supervision framework. [ABSTRACT FROM AUTHOR]

As an education model that focuses on the cultivation of students' creativity and practical ability, the integration of maker education and the development of artificial intelligence is of great significance. In order to analyze and study the application effect of artificial intelligence in maker education, based on this, this study adopts the method of questionnaire and case analysis to investigate and analyze students from three dimensions: learning experience, learning attitude and learning emotion, learning effect and transfer. The survey results show that, on the whole, the integration of artificial intelligence technology into maker education can significantly improve students' learning feelings and attitudes, and enhance the enthusiasm of learning emotions. The research also finds that it has a positive impact on students' learning effect and transfer ability. Learning experience, learning attitude and learning emotion, learning effect and learning transfer are related to gender, teaching time, teaching subjects or majors. Therefore, the article will suggest creating a maker cultural atmosphere, creating a maker resource space, and improving the maker education system. This provides a valuable reference for further promoting the development of maker education courses in the field of artificial intelligence. [ABSTRACT FROM AUTHOR]

Background: Repeated sessions of training and non-invasive brain stimulation have the potential to enhance cognition in patients with cognitive impairment. We hypothesized that combining cognitive training with anodal transcranial direct current stimulation (tDCS) will lead to performance improvement in the trained task and yield transfer to non-trained tasks. Methods: In our randomized, sham-controlled, double-blind study, 46 patients with cognitive impairment (60–80 years) were randomly assigned to one of two interventional groups. We administered a 9-session cognitive training (consisting of a letter updating and a Markov decision-making task) over 3 weeks with concurrent 1-mA anodal tDCS over the left dorsolateral prefrontal cortex (20 min in tDCS, 30 s in sham group). Primary outcome was trained task performance (letter updating task) immediately after training. Secondary outcomes included performance in tasks testing working memory (N-back task), decision-making (Wiener Matrices test) and verbal memory (verbal learning and memory test), and resting-state functional connectivity (FC). Tasks were administered at baseline, at post-assessment, and at 1- and 7-month follow-ups (FU). MRI was conducted at baseline and 7-month FU. Thirty-nine participants (85%) successfully completed the intervention. Data analyses are reported on the intention-to-treat (ITT) and the per-protocol (PP) sample. Results: For the primary outcome, no difference was observed in the ITT (β = 0.1, 95%-CI [− 1.2, 1.3, p = 0.93] or PP sample (β = − 0.2, 95%-CI [− 1.6, 1.2], p = 0.77). However, secondary analyses in the N-back working memory task showed that, only in the PP sample, the tDCS outperformed the sham group (PP: % correct, β = 5.0, 95%-CI [− 0.1, 10.2], p = 0.06, d-prime β = 0.2, 95%-CI [0.0, 0.4], p = 0.02; ITT: % correct, β = 3.0, 95%-CI [− 3.9, 9.9], p = 0.39, d-prime β = 0.1, 95%-CI [− 0.1, 0.3], p = 0.5). Frontoparietal network FC was increased from baseline to 7-month FU in the tDCS compared to the sham group (pFDR < 0.05). Exploratory analyses showed a correlation between individual memory improvements and higher electric field magnitudes induced by tDCS (ρtDCS = 0.59, p = 0.02). Adverse events did not differ between groups, questionnaires indicated successful blinding (incidence rate ratio, 1.1, 95%-CI [0.5, 2.2]). Conclusions: In sum, cognitive training with concurrent brain stimulation, compared to cognitive training with sham stimulation, did not lead to superior performance enhancements in patients with cognitive impairment. However, we observed transferred working memory benefits in patients who underwent the full 3-week intervention. MRI data pointed toward a potential intervention-induced modulation of neural network dynamics. A link between individual performance gains and electric fields suggested dosage-dependent effects of brain stimulation. Together, our findings do not support the immediate benefit of the combined intervention on the trained function, but provide exploratory evidence for transfer effects on working memory in patients with cognitive impairment. Future research needs to explore whether individualized protocols for both training and stimulation parameters might further enhance treatment gains. Trial registration: The study is registered on ClinicalTrials.gov (NCT04265378). Registered on 7 February 2020. Retrospectively registered. [ABSTRACT FROM AUTHOR]

Numerous studies have found that the Bayesian framework, which formulates the optimal integration of the knowledge of the world (i.e. prior) and current sensory evidence (i.e. likelihood), captures human behaviours sufficiently well. However, there are debates regarding whether humans use precise but cognitively demanding Bayesian computations for behaviours. Across two studies, we trained participants to estimate hidden locations of a target drawn from priors with different levels of uncertainty. In each trial, scattered dots provided noisy likelihood information about the target location. Participants showed that they learned the priors and combined prior and likelihood information to infer target locations in a Bayes fashion. We then introduced a transfer condition presenting a trained prior and a likelihood that has never been put together during training. How well participants integrate this novel likelihood with their learned prior is an indicator of whether participants perform Bayesian computations. In one study, participants experienced the newly introduced likelihood, which was paired with a different prior, during training. Participants changed likelihood weighting following expected directions although the degrees of change were significantly lower than Bayes-optimal predictions. In another group, the novel likelihoods were never used during training. We found people integrated a new likelihood within (interpolation) better than the one outside (extrapolation) the range of their previous learning experience and they were quantitatively Bayes-suboptimal in both. We replicated the findings of both studies in a validation dataset. Our results showed that Bayesian behaviours may not always be achieved by a full Bayesian computation. Future studies can apply our approach to different tasks to enhance the understanding of decision-making mechanisms. Author summary: Bayesian decision theory has emerged as a unified approach for capturing a wide range of behaviours under uncertainty. However, behavioural evidence supporting that humans use explicit Bayesian computation is scarce. While it has been argued that knowledge generalization should be treated as hard evidence of the use of Bayesian strategies, results from previous work were inconclusive. Here, we develop a marker that effectively quantifies how well humans transfer learned priors to a new scenario. Our marker can be applied to various tasks and thus can provide a path forwarding the understanding of psychological and biological underpinnings of inferential behaviours. [ABSTRACT FROM AUTHOR]

The introduction of virtual reality (VR) to flying training has recently gained much attention, with numerous VR companies, such as VRM Switzerland and VRpilot, looking to enhance the training process. Such a considerable change to how pilots are trained is a subject that warrants careful consideration. Examining the effect that VR has on learning in other areas gives us an idea of how VR can be suitably applied to flying training. Some of the benefits offered by VR include increased safety, decreased costs, accelerated learning process, and increased environmental sustainability. Nevertheless, some challenges ahead for developers to consider are negative transfer of learning, cybersickness, and failure for users to adopt the technology. In addition to this discussion, existing technologies are presented and compared across a number of areas. Future directions for research and development in VR flying training are considered, highlighting the importance of thorough testing procedures, and shifting research focus to the considerate integration of VR into flying training procedures. [ABSTRACT FROM AUTHOR]

Last train timetable plays an important role in actual operation of metro systems, especially in big cities. Appreciable transfer passengers rely on last train connections to reach their destinations, which requires a coordinated last train timetable. This study is devoted to dealing with the last train timetabling problem in metro networks, and proposes an integer linear programing model to coordinate last train connections at transfer stations to improve the transfer accessibility in terms of served transfer passengers and effective last train connections. Several measures at the planning level are used to guarantee feasibility of the optimized last train timetable when applied in practice. Then, the schedule-based transfer network and pre-process method are presented to shrink the scale of binary variables, which can expedite the solution process when solving problems in large-scale networks. Finally, some experiments based on the Qingdao metro network are carried out to verify the effectiveness of the timetabling method, and results show that dwell time and service ending time have direct influence on network accessibility. Comparison analyses show that the proposed model is able to generate a coordinated enough last train timetable with higher transfer accessibility and better practicality, which can provide support for operators to improve the last train service in a metro network. [ABSTRACT FROM AUTHOR]

Objective: The purpose of this meta-analytic review is to examine the relationship between increased psychological pressure and Use of Force (UOF) behaviours, identifying current training methodologies and effectiveness of transfer of training interventions in high threat-simulated scenarios. Background: Data from UOF performance within Law Enforcement indicates a low transfer of marksmanship training into real-world UOF, resulting in unnecessary damage to property, personal injury and increased risk to loss of life. This meta-analysis examines both the impact of increased pressure and current training interventions. Method: A meta-analysis was conducted across a wide range of published research to answer the primary research questions. Results: Increased levels of perceived pressure demonstrated an average decrease in marksmanship accuracy of 14.8%, together with a small increase in incorrect Decision Making (DM) and faster reaction Times (RT). Experience demonstrated a mitigating effect for pressure for marksmanship with a 1.1% increase for every one year of service but no effect on DM or RT. Training interventions utilizing a variety of early contextually relevant exposures to increased pressure improved performance over traditional training on average by 10.6%. Conclusion: The outcomes illustrate the negative effect of pressure on marksmanship and UOF behaviours, and that early exposure to contextually relevant pressure may increase the transfer of training to real-world performance. Application: Occupational experience is an important component in reducing the impact of pressure on UOF performance, and transfer of training may be enhanced through training methodologies that combine early exposure to contextually relevant pressure, that may replicate the benefits of experience. [ABSTRACT FROM AUTHOR]

Introduction: The mismatch between training and testing data distribution causes significant degradation in the deep learning model performance in multi-ethnic scenarios. To reduce the performance differences between ethnic groups and image domains, we built a deep transfer learning model with adaptation training to predict uncorrected refractive errors using posterior segment optical coherence tomography (OCT) images of the macula and optic nerve. Methods: Observational, cross-sectional, multicenter study design. We pre-trained a deep learning model on OCT images from the B&VIIT Eye Center (Seoul, South Korea) (N = 2602 eyes of 1301 patients). OCT images from Poona Eye Care (Pune, India) were chronologically sorted into adaptation training data (N = 60 eyes of 30 patients) for transfer learning and test data (N = 142 eyes of 71 patients) for validation. Deep learning models were trained to predict spherical equivalent (SE) and mean keratometry (K) values via transfer learning for domain adaptation. Results: Both adaptation models for SE and K were significantly better than those without adaptation (P < 0.001). In myopia/hyperopia classification, the model trained on circular optic disc OCT images yielded the best performance (accuracy = 74.7%). It also performed best in estimating SE with the lowest mean absolute error (MAE) of 1.58 D. For classifying the degree of corneal curvature, the optic nerve vertical algorithm performed best (accuracy = 65.7%). The optic nerve horizontal model achieved the lowest MAE (1.85 D) when predicting the K value. Saliency maps frequently highlighted the retinal nerve fiber layers. Conclusions: Adaptation training via transfer learning is an effective technique for estimating refractive errors and K values using macular and optic nerve OCT images from ethnically heterogeneous populations. Further studies with larger sample sizes and various data sources are needed to confirm the feasibility of the proposed algorithm. [ABSTRACT FROM AUTHOR]

This paper focuses on the application of meta-reinforcement learning to the robust design of low-thrust interplanetary trajectories in the presence of multiple uncertainties. A closed-loop control policy is used to optimally steer the spacecraft to a final target state despite the considered perturbations. The control policy is approximated by a deep recurrent neural network, trained by policy-gradient reinforcement learning on a collection of environments featuring mixed sources of uncertainty, namely dynamic uncertainty and control execution errors. The recurrent network is able to build an internal representation of the distribution of environments, thus better adapting the control to the different stochastic scenarios. The results in terms of optimality, constraint handling, and robustness on a fuel-optimal low-thrust transfer between Earth and Mars are compared with those obtained via a traditional reinforcement learning approach based on a feed-forward neural network. • In interplanetary space missions, the spacecraft trajectory is affected by multiple uncertainties. • Meta-reinforcement learning can be applied seamlessly to any uncertainty and dynamic model. • A recurrent neural network is used as a history-dependent closed-loop control policy. • The network is trained on a low-thrust transfer featuring dynamic uncertainties and control execution errors. • Meta-reinforcement learning shows improved performance and robustness compared to standard reinforcement learning. [ABSTRACT FROM AUTHOR]

Skill training aims to improve the performance of the task at hand and aims to transfer the acquired skill to related tasks. Both skill training and skill transfer are part of our everyday lives, and essential for survival, and their importance is reflected in years of research. Despite these enormous efforts, however, the complex relationship between skill training and skill transfer is not yet portrayed completely. Building upon two theories, we probed this relationship through the example of bimanual learning with a large cross‐sectional design (N = 450) using an online framework. We designed five training tasks which differed in the variance of the training material (schema theory) and three transfer tasks differing in their similarity to the training task (identical elements theory). Theoretically, the five training tasks and the three transfer tasks varied approximately linearly from each other. Empirical data, however, suggested merely the presence of three statistically different training tasks and two significantly different transfer tasks, indicating a nonlinear relationship. Against our expectation, Bayesian statistics suggested that the type of skill training was not related to the type of skill transfer. However, the amount of skill training was positively related to the amount of skill transfer. Together, we showed that motor learning studies can be conducted online. Further, our results shed light on the complex relationship between skill training and skill transfer. Understanding this relationship has wide‐ranging practical implications for the general population, particularly for musicians, athletes and patients recovering from injury. [ABSTRACT FROM AUTHOR]