Learning Stage Research Articles

A data-centric approach for ethical and trustworthy AI in journalism

AI-driven journalism refers to various methods and tools for gathering, verifying, producing, and distributing news information. Their potential is to extend human capabilities and create new forms of augmented journalism. Although scholars agreed on the necessity to embed journalistic values in these systems to make AI systems accountable, less attention was paid to data quality, while the results’ accuracy and efficiency depend on high-quality data in any machine learning task. Assessing data quality in the context of AI-driven journalism requires a broader and interdisciplinary approach, relying on the challenges of data quality in machine learning and the ethical challenges of using machine learning in journalism. To better identify these, we propose a data quality assessment framework to support the collection and pre-processing stages in machine learning. It relies on three of the core principles of ethical journalism—accuracy, fairness, and transparency—and participates in the shift from model-centric to data-centric AI, by focusing on data quality to reduce reliance on large datasets with errors, making data labelling consistent, and better integrating journalistic knowledge.

Multi-agent system-based framework for an intelligent management of competency building

To measure the effectiveness of learning activities, intensive research works have focused on the process of competency building through the identification of learning stages as well as the setup of related key performance indictors to measure the attainment of specific learning objectives. To organize the learning activities as per the background and skills of each learner, individual learning styles have been identified and measured by several researchers. Despite their importance in personalizing the learning activities, these styles are difficult to implement for large groups of learners. They have also been rarely correlated with each specific learning stage. New approaches are, therefore, needed to intelligently coordinate all the learning activities while self-adapting to the ongoing progress of learning as well as to the specific requirements and backgrounds of learners. To address these issues, we propose in this paper a new framework for an intelligent management of the competency building process during learning. Our framework is based on a recursive spiral Assess-Predict-Oversee-Transit model that is orchestrated by a multi-agent system. This system is particularly responsible of enabling smart transitions between learning stages. It is also responsible of assessing and predicting the process of competency building of the learner and, then, making the right decisions about the learning progress, accordingly. Results of our solution were demonstrated via an Augmented Reality app that we created using the Unity3D engine to train learners on Air Conditioner maintenance.

An efficient approach for EMG controlled pattern recognition system based on MUAP identification and segregation

An Electromyography (EMG) based pattern recognition system constitutes various steps of signal processing and control engineering from signal acquisition to real-time control. Efficient control of external devices largely depends on the signal processing steps executed before the final output. This work presents a new approach to signal processing using Motor Unit Action Potential (MUAP) based signal decomposition and segmentation. An MUAP is a neurological response during muscle contraction. Due to the higher contact area of surface electrodes, MUAPs from multiple muscles are captured. An MUAP generated from a single muscle usually has identical waveshapes and similar discharging rates and usually lasts for 8–15 ms. These are known as primary MUAPs. The proposed algorithm identifies and uses the primary observed MUAPs for feature extraction and classification. Firstly, noise signals are eliminated by a determined noise margin, which also separates the active muscle movement signals. Next, a novel MUAP identification algorithm is implemented to detect the MUAP trains. Then, identified primary MUAPs are used to make segments with variable widths to extract feature vectors. Based on the correlation score of all the primary MUAPs, the segmentation is performed, which results in segmentation width varying from 110–200 ms. The achieved segmentation width is lesser than the conventional overlapping and non-overlapping methods — the proposed approach results in a 20 to 50% reduction in the segmentation width. Four different classifiers are tested during the machine learning stage to investigate the performance of the proposed approach. The obtained feature sets are then used to train the Linear Discriminant Analysis (LDA), K-Nearest Neighbor (kNN), Decision Tree (DT), and Random Forest (RF) classifiers. The classifiers are tested with precision, recall, F1 score, and accuracy. The kNN and DT classifiers performed better than the LDA and RF classifiers. The maximum precision and recall are 100% while the maximum achieved accuracy is 98.56%. The comparative results show higher accuracy even at lower segmentation widths than the conventional constant window scheme. The kNN and DT classifiers provide a 5% to 15% increment in accuracy compared to the constant window segmentation-based approach.

11-month-olds recognize the teacher–student relationship

Understanding how preverbal infants perceive and differentiate the roles of teachers and students can provide profound insights into the early stages of human learning and cognition. We thus developed a paradigm based on the task order and the interactive position to determine if infants understand teacher–student relationships. Five experiments (N = 104), reveal that, after watching the teacher agent instruct the student agent from failure to success, 11-month-old infants looked at the teacher for a longer time when they watched a new naïve agent seek information in a new ambiguous scenario. Three additional control experiments excluded alternative interpretations of infants’ selective looking, suggesting that agents’ order-related actions influenced infants’ identification. Another control experiment demonstrated the importance of the position when the teacher and the student had interactive exchanges. These findings indicate that preverbal infants can recognize the teacher–student relationship, which is crucial to accessing effective information in early life.

Beyond transfer learning: Leveraging ancillary images in automated classification of plankton

AbstractWe assess whether a supervised machine learning algorithm, specifically a convolutional neural network (CNN), achieves higher accuracy on planktonic image classification when including non‐plankton and ancillary plankton during the training procedure. We focus on the case of optimizing the CNN for a single planktonic image source, while considering ancillary images to be plankton images from other instruments. We conducted two sets of experiments with three different types of plankton images (from a Zooglider, Underwater Vision Profiler 5, and Zooscan), and our results held across all three image types. First, we considered whether single‐stage transfer learning using non‐plankton images was beneficial. For this assessment, we used ImageNet images and the 2015 ImageNet contest‐winning model, ResNet‐152. We found increased accuracy using a ResNet‐152 model pretrained on ImageNet, provided the entire network was retrained rather than retraining only the fully connected layers. Next, we combined all three plankton image types into a single dataset with 3.3 million images (despite their differences in contrast, resolution, and pixel pitch) and conducted a multistage transfer learning assessment. We executed a transfer learning stage from ImageNet to the merged ancillary plankton dataset, then a second transfer learning stage from that merged plankton model to a single instrument dataset. We found that multistage transfer learning resulted in additional accuracy gains. These results should have generality for other image classification tasks.

The Early Reading and Music partnership

There is broad agreement, in relation to the high gravitas of ‘the big six’ in teaching early reading, the pillars of effective reading instruction (Moats, 1999; NRP, 2000; Konza, 2014). Notwithstanding, there exists a body of evidence that indicates the arts, in particular music, enhance and motivate other learning, helping young children to become readers. This suggests the explicit teaching of early reading and music together can yield powerful learning for students, the conflation of which improves and accelerates both reading and music mastery. This research investigates whether participation in music instruction has beneficial effects on young children’s learning and acquisition of early reading skills; specifically, if a link exists between early reading and music acquisition. There are three parts to this research: a) literature review; b) email surveys/ semi structured interviews; c) classroom observations. Results suggest that music serves as a natural bridge to literacy, strengthening auditory processing, and scaffolding learning phonemes and graphemes. Music aids memory because the beat, melody and harmony carry semantic content, clarifying meaning. Children learn with less effort and remember information more easily when rhyme and rhythm are present. Nursery rhymes, poems and action songs engage children through playing with language, suggesting that a transdisciplinary pedagogical approach is a potential solution to challenges currently faced in education in relation to disengagement from traditional schooling and declining reading levels. An early reading and music partnership is an approach to learning that may benefit young learners along the road to becoming readers.

Challenges in Promoting Learner Autonomy in Blended Learning: Perspectives from English as a Foreign Language Teachers in China

Blended learning is gaining popularity as a prevailing practice in higher education worldwide. This study explores the challenges faced by experienced English as a Foreign Language teachers in promoting learner autonomy within blended learning environments. Drawing from interviews and classroom observations conducted with five EFL teachers at a private university in China, the study identifies four key challenges: technological, pedagogical, cultural, and institutional. Technological challenges revolve around concerns regarding technical support, internet reliability, and hardware availability, highlighting the need for sturdy technical infrastructure and supportive mechanisms. Pedagogical challenges include designing effective blended learning activities that balance teacher-led instruction with opportunities for student autonomy, emphasizing the importance of scaffolding learning tasks, and providing adequate support for independent learning. Cultural challenges underscore the influence of cultural factors on student attitudes towards autonomy and authority in the classroom, emphasizing the necessity for culturally responsive teaching practices. Institutional challenges highlight barriers stemming from rigid curriculum requirements, limited professional development opportunities, and a lack of administrative support, underscoring the importance of systemic support and organizational change. The findings offer valuable guidance for institutions, administrators, and EFL teachers on how to design more effective blended EFL courses. Limitations and suggestions for future studies such as larger sample size, rigorous data collection method, and longitudinal studies are provided to enhance our understanding of promoting learner autonomy in blended learning environments.

Scaffolding learning: From specific to generic with large language models

Scaffolding learning: From specific to generic with large language models

Implementasi Augmented Reality untuk Pengenalan Bangun Ruang

Solid Geometry is a part of mathematics learning that has various forms and types. The diversity of spatial shapes has area and volume formulas so that you can know the number of sides and area of each spatial shape. By utilizing Augmented Reality technology to introduce spatial structures, it makes it interesting to study. The aim of this research is to create an AR application for solid geometry in an Android-based 3D model so that it can facilitate learning in a fun way. This AR application was created using Vuforia, Unity and a marker to be able to call up the space that has been created. From this research, it is hoped that the interim results can help the learning and teaching process on spatial building materials. Application research results that can be used for the learning stage.

BUILDING STEM TOPICS IN BLENDED LEARNING ABOUT ENERGY TO CONTRIBUTE TO SUSTAINABLE DEVELOPMENT EDUCATION

With the goal of education aimed at sustainable development as reflected in the 2018 General Education Program, schools in general and the subject of physics in particular need to educate students to raise awareness about energy issues and how to use energy efficiently, safely, and reasonably. Therefore, this report presents the research results on building the STEM topic “Energy” in blended learning, contributing to sustainable development education. The research results propose a process of blended learning stages and the development of teaching materials to support teachers in teaching STEM topics about energy, helping students develop competencies and the ability to cope with future challenges. Keywords: STEM education; energy; blended learning; sustainable development education.

Class incremental learning with analytic learning for hyperspectral image classification

Hyperspectral image classification (HSIC) is crucial for applications in agriculture and environmental monitoring. As ground objects evolve and remote sensing technology progresses, there is a growing need for HSIC models that can adapt to new data classes without requiring to be retrained from scratch. Throughout the continual learning procedure, the model is expected to not only effectively extract spatial–spectral features from the hyperspectral image but also alleviate the issue of catastrophic forgetting, i.e., the model forgets the learned classes’ knowledge when accessing novel classes during the training process. In this paper, for the HSIC task, we propose a class incremental learning method (HSI-CIL) that is based on analytic learning, a technique that converts network training into linear problems. Specifically, The HSI-CIL model consists of a lightweight feature extractor, a Feature Processing Module (FPM) and an Analytic Linear Classifier (ALC). This model does not need data storage for old and new classes and has only one epoch in the incremental learning stage, so it has lower consumption of resources and training time than several attempts have been proposed for addressing catastrophic forgetting. We perform abundant experiments with the proposed HSI-CIL on three publicly available hyperspectral datasets including Indian Pines, Pavia University, and Salinas. The experiments demonstrate that our HSI-CIL exceeds the state-of-the-art class incremental learning (CIL) techniques applied in HSIC with a certain gap.

Generative domain-adapted adversarial auto-encoder model for enhanced ultrasonic imaging applications

In this study, we propose a class-conditioned Generative Adversarial Autoencoder (cGAAE) to improve the realism of simulated ultrasonic imaging techniques, in particular the Multi-modal Total Focusing Method (M-TFM), based on the availability of both simulated and experimental TFM images. In particular, this work studied the case of the inspection of a complex geometry block representative of weld-inspection problem based on ultrasonic multi-elements probe. The cGAAE is represented by a tailored learning schema, trained in a semi-supervised fashion on a labeled mixture of synthetic (class 0) and experimental (class 1) M-TFM images, obtained under different meaningful inspection set-ups parameters (i.e., the celerity of the transverse ultrasonic wave, the specimen back-wall slope and height, the flaw tilt and heigh). That is, the cGAAE schema consists in a combination of learning stages involving class-conditioned spatial-transformers and arbitrary style transfer endows the cGAAE of powerful generative features, such as quasi real-time generation of M-TFM images by sweep of the inspection parameters. We exploited the cGAAE model to improve the realism of simulated M-TFM images and enhance the accuracy of the inverse problem, aiming at estimating the inspection parameters based on experimental acquisitions.

Exploration on the Teaching Practice of Color Foundation of Painting Majors in Colleges and Universities

Color foundation teaching is the main content of the color expression ability in the painting major of colleges and universities. The traditional color training methods are prone to being misunderstood as the cultivation of the representation techniques of objects of varying difficulties, thereby neglecting that the fundamental purpose of color foundation training is for creative expression. Compared with the traditional course model, the creative consciousness and concept guidance model is a reliable approach to address the expression issues during students' color foundation training at all stages. The traditional color teaching method of combining sketching from life and silent drawing often leads students to form passive, indiscriminate, and emotionless simulations of objects, causing the objects themselves to lose their significance of being depicted. The color lacks directionality and expressive purpose. Works overly rely on the accidental effects produced by unconscious object projection or the meticulousness of work processing, resulting in "sketching" becoming "stifled sketching". The creative consciousness and concept guidance model emphasizes extracting the characteristics of objects. Educators attach importance to discovering and affirming the accidental writing and incompleteness that occur during students' color practice, helping students create their own color imagery based on the objects they are representing at any stage of color learning.

English Vowel Perception in Spanish-English Bilingual Preschoolers: Multiple-Talker Input Is Only Beneficial for Children With High Language Exposure Levels.

This study examines English vowel perception in Spanish-English bilingual preschoolers, comparing bilinguals' perception patterns to those of monolinguals and examining how child internal (age) and external variables (input quantity and input diversity) predict perceptual performance. Sixty children between 3;6 and 5;6 (years;months) of age participated in the study, 28 of whom were Spanish-English bilinguals and 32 English monolinguals. Perception was assessed through a forced-choice minimal-pair identification task in which children heard synthesized audio stimuli (i.e., "sheep" and "ship") that varied systematically along the /i-ɪ/ continuum and were asked to match them with one of two pictures. The data were analyzed with Bayesian mixed-effects logistic regression analyses, modeling responses as a function of continuum step, language background (monolingual or bilingual), age, English exposure (i.e., input quantity), and number of English input providers (i.e., input diversity). The results indicate that, despite displaying nonnative English stop voicing perception in a previous study, the same bilingual children showed English /i-ɪ/ perception patterns that did not differ from those of monolinguals. While age did not predict vowel perception, input quantity and diversity jointly interacted to moderate how well children perceived the /i-ɪ/ contrast. Specifically, diverse input promoted perceptual performance in children who received high levels of English exposure-and who presumably had more advanced English language skills, whereas it limited perceptual performance in children with more limited English exposure and skills. This study shows that bilingual children can show monolingual-like perception patterns for some sounds while displaying nontarget perception for others. This is the first study to demonstrate that language exposure mediates the role of input diversity on speech sound development, suggesting that varied input can be more or less beneficial for speech sound development based on the learner's language learning stage.

중국내 한국어 학습자의 문법용어 이해도에 대한 통계적 분석

Objectives The purpose of this paper is to do a quantitative survey of the Intelligibility of Korean Grammatical Terms for Chinese Learner. Methods In this study, undergraduate and graduate students in China were surveyed to determine the level of understanding of Korean grammatical terms among Chinese learners. The SPSS statistical program is used to find out the statistical skills of Chinese learners' comprehension of grammatical terms and the differences in comprehension by learning stage. Results Based on the results of the analysis so far, the hypothesis that there is a difference in the comprehension of grammatical terms by learning stage is rejected, i.e., there is a significant difference in the individual grammatical term items, but the t-test on the mean does not show a significant difference. Conclusions The results of measuring the comprehension of grammatical terms using SPSS were not statistically significant at the 0.05 level of significance for each undergraduate level.

Purkinje cells in Crus I and II encode the visual stimulus and the impending choice as monkeys learn a reinforcement based visuomotor association task.

Visuomotor association involves linking an arbitrary visual cue to a well-learned movement. Transient inactivation of Crus I/II impairs primates' ability to learn new associations and delays motor responses without affecting the kinematics of the movement. The simple spikes of Purkinje cells in the Crus regions signal cognitive errors as monkeys learn to associate specific fractal stimuli with movements of the left or right hand. Here we show that as learning progresses, the simple spike activity of individual neurons becomes more selective for stimulus-response associations, with selectivity developing closer to the appearance of visual stimuli. Initially, most neurons respond to both associations, irrespective of the identity of the stimulus and the associated movement, but as learning advances, more neurons distinguish between specific stimulus-hand associations. Using a linear decoder, it was found that in early learning stages, the visual stimulus can be decoded only when the choice can also be decoded. As learning improves, the visual stimulus is decoded earlier than the choice. A simple model can replicate the observed simple spike signals and the monkeys' behavior in both the early and late learning stages.

Applying a community-engaged participatory machine learning model.

Although predictive algorithms have been described as the definitive solution to bias in health care, machine learning techniques may also propagate existing health inequities within the community context. However, there may be ways in which machine learning techniques can help community psychologists, public health researchers and practitioners identify patterns in data in a way that empowers improved outcomes. Incorporating community insight in all stages of machine learning research mitigates bias by positioning members of underrepresented communities as the experts of their lived experiences. As community psychologists already prioritize community-based participatory practices, we propose three core guiding principles for a community-engaged participatory model for research using machine learning techniques: shared decision-making, reflexivity and structural humility, and flexibility and adaptability. Guided by these three principles, we emphasize grounding priority setting, problem formation, model assumptions, and interpretation of the resulting algorithmic patterns in the truths born from the lived experiences of people closest to the problem. We also suggest opportunities for bidirectional and mutually empowering partnerships between algorithmic scientists and the communities to which their algorithms will be applied. Inclusion of community stakeholders in all stages of machine learning for health research provides an opportunity to develop algorithms that are both highly effective and ethically grounded in the lived experiences of target populations.

Prospective teachers’ views and experiences with e-portfolios

ABSTRACT The use of electronic portfolios (e-portfolios) is a burgeoning area of research, with many teacher education programs utilising as both a product of and a process of reflective practice. This case study aimed to explore prospective primary teachers’ (students) perceptions, perspectives and experiences of the use of e-portfolios, focusing on the affordances and limitations of using e-portfolios in Initial Teacher Education (ITE). Eleven second-year undergraduate students from secondary and primary ITE programs were interviewed. A reflexive thematic analysis was conducted on the transcripts, in which the module tutors’ experiences were also incorporated. Findings suggest that e-portfolios can be valuable agentic and reflective pedagogical tools that can scaffold learning, although with a number of important caveats for future implementations. This paper speaks to the paucity of research that extends the view of e-portfolios as simply repositories for content.

The Effect of Age of First Exposure on Vocabulary, Mean Length of Utterance, Morphosyntactic Accuracy, and Semantic and Sentence-Level Patterns in the First 2 Years of French Second-Language Learning by Preschool- to Adolescent-Age Mandarin Speakers.

Bilingual assessment is particularly difficult in the very first period of children's second language (L2) exposure. This exploratory, longitudinal study examined L2 learning after 1 and 2 years of L2 exposure by young immigrants and how it is affected by their age at first exposure to the L2 (AoE). Participants were 18 immigrants ranging in age from 2;11 to 14;2 (years;months), all within their first year in Montreal at Time 1, enrolled in a French school or day care, and from a Mandarin first language background. Participants were tested again a year later. Measures included receptive and expressive French vocabulary tests and conversational language samples analyzed using traditional measures of mean length of utterance (MLU) and morphological accuracy as well as novel measures of semantic and sentence-level patterns. Performance was relatively high already at Time 1 and increased significantly at Time 2 in both vocabulary and MLU. At Time 2, vocabulary scores were below normative values, whereas MLU was within expected values relative to monolingual and simultaneous bilinguals for the majority of the participants. However, higher MLUs were accompanied by more instances of both semantic errors and creative semantic strategies. French performance was strongly related to AoE; with amount of exposure equivalent, older participants outperformed the younger ones on MLU and vocabulary. Semantic errors and creative uses were strongly predicted by AoE; however, morphological accuracy and number of agrammatical utterances were not. This initial period of French learning involved a rapid growth spurt for most of the participants. We argue that the pattern observed, particularly among the older children, constitutes an early stage of L2 learning characterized by long utterances that are also frequently hard to understand as speakers encounter challenges and use creative strategies in their attempt to convey meaning. Comparison with normative reference bases for monolinguals and bilinguals with greater cumulative L2 exposure who have similar MLUs should be done with much caution during this early period.

Underwater variable zoom: Depth-guided perception network for underwater image enhancement

The underwater images often suffer from color deviations and blurred details. To address these issues, many methods employ networks with an encoder/decoder structure to enhance the images. However, the direct skip connection overlooks the differences between pre- and post-features, and deep network learning introduces information loss. This paper presents an underwater image enhancement network that focuses on pre-post differences. The network utilizes a multi-scale input and output framework to facilitate the underwater image enhancement process. A novel cross-wise transformer module (CTM) is introduced to guide the interactive learning of features from different periods, thereby enhancing the emphasis on detail-degraded regions. To compensate for the information loss within the deep network, a feature supplement module (FSM) is devised for each learning stage. FSM merges the multi-scale input features, effectively enhancing the visibility of underwater images. Experimental results across several datasets demonstrate that the integrated modules yield significant enhancements in network performance. The proposed network exhibits outstanding performance in both visual comparisons and quantitative metrics. Furthermore, the network also exhibits good adaptability in additional visual tasks without the need for parameter tuning. The code and results are released in https://github.com/WindySprint/UVZ.