Learning Task Research Articles

Rapid Classification of Sugarcane Nodes and Internodes Using Near-Infrared Spectroscopy and Machine Learning Techniques

Accurate and rapid discrimination between nodes and internodes in sugarcane is vital for automating planting processes, particularly for minimizing bud damage and optimizing planting material quality. This study investigates the potential of visible-shortwave near-infrared (Vis–SWNIR) spectroscopy (400–1000 nm) combined with machine learning for this classification task. Spectral data were acquired from the sugarcane cultivar Khon Kaen 3 at multiple orientations, and various preprocessing techniques were employed to enhance spectral features. Three machine learning algorithms, linear discriminant analysis (LDA), K-Nearest Neighbors (KNNs), and artificial neural networks (ANNs), were evaluated for their classification performance. The results demonstrated high accuracy across all models, with ANN coupled with derivative preprocessing achieving an F1-score of 0.93 on both calibration and validation datasets, and 0.92 on an independent test set. This study underscores the feasibility of Vis–SWNIR spectroscopy and machine learning for rapid and precise node/internode classification, paving the way for automation in sugarcane billet preparation and other precision agriculture applications.

Teaching Science Outdoors: Supporting Pre-Service Teachers’ Skill Development with the Help of Available Mobile Applications

Outdoor environments provide excellent teaching and learning experiences in science education. However, many teachers find outdoor teaching challenging. In this study, we investigated factors supporting skill development and learning among pre-service teacher during a blended science didactics course that includes mobile interaction in outdoor environments. Available WhatsApp mobile application was used as an interaction platform between the pre-service teachers’ and the teacher educator. Based on the findings, the pre-service teachers learned easy ways of using outdoor environments with pupils. They also identified challenges that may arise in outdoor teaching and upskilled on how to overcome them. From the perspective of interaction, submitting learning tasks, especially visual observations, through mobile messaging and reviewing tasks of other students in the application were perceived as important. However, the most crucial benefit of mobile interaction was considered to be the teacher’s real-time feedback.

Humanoid Robot-Assisted Learning for Level 1 Autistic Children: Developing Test Plans to Enhance Number Recognition Skills

Autism Spectrum Disorder (ASD) impacts the social, communication, and cognitive development of affected individuals. Level 1 autism, the mildest form of ASD, presents unique challenges in the acquisition of crucial skills such as number recognition, which serves as a foundational aspect of children's cognitive development. Recent studies have shown that humanoid robots can offer effective intervention strategies for children with autism, fostering learning in a structured, engaging, and interactive environment. This study aims to develop and evaluate test plans using humanoid robots to enhance number recognition skills in Level 1 children with autism. Based on the findings, a test plan was designed to address the unique learning needs of Level 1 children with autism in number recognition. Participants were selected based on specific criteria, ensuring a diverse and representative sample. The humanoid robot's capabilities and features were detailed, emphasizing its role in facilitating the test plan. The research methodology outlined the test plan components, implementation procedure, data collection, and analysis techniques. Results demonstrated that the humanoid robot-assisted test plan was effective in enhancing number recognition skills among Level 1 children with autism. The participants exhibited significant improvements in their abilities to identify and manipulate numbers, as well as increased engagement and motivation in learning tasks. The study also revealed that the interactive and supportive nature of the humanoid robot fostered a positive learning environment, which contributed to the observed improvements in number recognition skills. The implications of these findings extend to the broader field of autism intervention, highlighting the potential benefits of incorporating humanoid robots into educational programs for children with autism.

Using Four-Component Instructional Design to Create an Interactive Point-of-Care Ultrasound Curriculum for Physician Associate Students.

Many physician assistant/associate (PA) programs lack point-of-care ultrasound (POCUS) education in the PA curricula and a standardized approach to ultrasound training. The four-component instructional design (4C/ID) model merges 4 concepts for developing effective instructional design content for complex content delivery, such as ultrasound. This research study created an interactive, 2-part ultrasound curriculum with an instructional designer for PA students using the 4C/ID model. Two interactive ultrasound curriculum models were created from the 4C/ID model using a knowledge-based precurriculum quiz, learning tasks with online modules and recordings, part-task practice with ultrasound experts, skills checkoffs, and a postcurriculum quiz including questions on students' perspectives of ultrasound. Two-tailed paired t-test analysis assessed the effectiveness of training. Evaluation of prelearning and postlearning quizzes demonstrated multiple statistically significant assessments supporting the use of instructional design technology for a PA program's ultrasound curriculum. This small study demonstrated that the 4C/ID model may be beneficial for PA POCUS training. The 2-part interactive curriculum improved PA student competency of ultrasound principles and applications and increased confidence for future clinical use of ultrasound.

Immersive virtual reality for learning exoskeleton-like virtual walking: a feasibility study

PurposeVirtual Reality (VR) has proven to be an effective tool for motor (re)learning. Furthermore, with the current commercialization of low-cost head-mounted displays (HMDs), immersive virtual reality (IVR) has become a viable rehabilitation tool. Nonetheless, it is still an open question how immersive virtual environments should be designed to enhance motor learning, especially to support the learning of complex motor tasks. An example of such a complex task is triggering steps while wearing lower-limb exoskeletons as it requires the learning of several sub-tasks, e.g., shifting the weight from one leg to the other, keeping the trunk upright, and initiating steps. This study aims to find the necessary elements in VR to promote motor learning of complex virtual gait tasks.MethodsIn this study, we developed an HMD-IVR-based system for training to control wearable lower-limb exoskeletons for people with sensorimotor disorders. The system simulates a virtual walking task of an avatar resembling the sub-tasks needed to trigger steps with an exoskeleton. We ran an experiment with forty healthy participants to investigate the effects of first- (1PP) vs. third-person perspective (3PP) and the provision (or not) of concurrent visual feedback of participants’ movements on the walking performance – namely number of steps, trunk inclination, and stride length –, as well as the effects on embodiment, usability, cybersickness, and perceived workload.ResultsWe found that all participants learned to execute the virtual walking task. However, no clear interaction of perspective and visual feedback improved the learning of all sub-tasks concurrently. Instead, the key seems to lie in selecting the appropriate perspective and visual feedback for each sub-task. Notably, participants embodied the avatar across all training modalities with low cybersickness levels. Still, participants’ cognitive load remained high, leading to marginally acceptable usability scores.ConclusionsOur findings suggest that to maximize learning, users should train sub-tasks sequentially using the most suitable combination of person’s perspective and visual feedback for each sub-task. This research offers valuable insights for future developments in IVR to support individuals with sensorimotor disorders in improving the learning of walking with wearable exoskeletons

Adding a Piece to the Puzzle: Children’s Exposure to Idioms

Idioms are figurative multiword expressions that need to be learned as part of the native phrasal vocabulary. While it has been shown that non-figurative multiword expressions are acquired with language exposure, the learning process for idioms may be different because the figurative meaning adds complexity to the learning task. Idiom vocabulary overall develops relatively late, but it is unknown to what extent children are exposed to idioms, and what kinds of idioms they encounter. Here, we investigated children’s idiom exposure and its effect on the development of idiom vocabulary in three studies: we explore the frequency of a well-tested set of Dutch idioms in a corpus of child literature, test idiom familiarity in a controlled setting in primary school children, and compare those findings to a set of online familiarity ratings. We find that children’s idiom exposure differs from adult idiom exposure, when comparing idiom frequencies based on children’s books and a corpus with resources for adults. Idiom decomposability and idiom frequencies from the children’s books, but not frequencies from the adult corpus, influenced the familiarity ratings of older children, suggesting that language exposure and idiom characteristics, such as decomposability, both play a role in idiom acquisition.

Network community detection via neural embeddings

Recent advances in machine learning research have produced powerful neural graph embedding methods, which learn useful, low-dimensional vector representations of network data. These neural methods for graph embedding excel in graph machine learning tasks and are now widely adopted. However, how and why these methods work—particularly how network structure gets encoded in the embedding—remain largely unexplained. Here, we show that node2vec—shallow, linear neural network—encodes communities into separable clusters better than random partitioning down to the information-theoretic detectability limit for the stochastic block models. We show that this is due to the equivalence between the embedding learned by node2vec and the spectral embedding via the eigenvectors of the symmetric normalized Laplacian matrix. Numerical simulations demonstrate that node2vec is capable of learning communities on sparse graphs generated by the stochastic blockmodel, as well as on sparse degree-heterogeneous networks. Our results highlight the features of graph neural networks that enable them to separate communities in the embedding space.

The pivotal role of clinical tutorial teaching in knowledge construction.

Students' new knowledge is gradually built up in the context of the task for which it is required and consolidated by applying it to clinical cases. As students see more and more clinical cases the knowledge emerges from an associative mesh of different levels of understanding. During tutorial clinical teaching, residents should be gradually exposed to an increasing range of real-world learning tasks and increasing levels of complexity. This exposure allows them to gradually develop shortcuts in the retrieval of their knowledge. This commentary provides a rationale for the construction of knowledge and the pivotal role that clinical tutorial teaching plays in this task.

PREDICTING STUDENT PERFORMANCE IN ONLINE LEARNING PLATFORMS: ANALYZING ENGAGEMENT METRICS WITH MACHINE LEARNING MODELS

The emergence of online learning platforms harbours both opportunities and challenges, this paper focuses on machine learning algorithms for deep diving into key predictors of students success in online environments, captured by engagement metrics such as quiz scores, participation in forums, and time spent on learning tasks. Data was collected through a survey of 50 students and then run through the Random Forest model. Major predictors of success include the assignment approach, quiz score, and motivation. The prediction from this model can be interesting because it can pick important features, but the overall accuracy is quite low at about 20%, which implies that larger datasets and more features might be necessary to enhance the predictions. Time management, motivation, and staying in touch are such variables that the research holds key to determining results for students. The results are important not only for online learning platforms but also for interventions suggested, which include early warning systems, personalised learning paths, and elements of gamification to support performance for improved student support at the right time for those most at risk. Future studies should consider more sophisticated machine learning models and the use of more objective performance data in their attempt to make more refined predictions.

Optogenetic activation of mesencephalic projections to the nucleus accumbens shell impairs probabilistic reversal learning by disrupting learning from negative reinforcement.

Cognitive flexibility, the capacity to adapt behaviour to changes in the environment, is impaired in a range of brain disorders, including schizophrenia and Parkinson's disease. Putative neural substrates of cognitive flexibility include mesencephalic pathways to the ventral striatum (VS) and dorsomedial striatum (DMS), hypothesized to encode learning signals needed to maximize rewarded outcomes during decision-making. However, it is unclear whether mesencephalic projections to the ventral and dorsal striatum are distinct in their contribution to flexible reward-related learning. Here, rats acquired a two-choice spatial probabilistic reversal learning (PRL) task, reinforced on an 80%|20% basis (correct|incorrect responses) that assessed the flexibility of behaviour to repeated reversals of response-outcome contingencies. We report that optogenetic stimulation of projections from the ventral tegmental area (VTA) to the nucleus accumbens shell (NAcS) in the VS significantly impaired reversal learning when optical stimulation was temporally aligned with negative feedback (i.e., reward omission). VTA → NAcS stimulation during other phases of the behavioural task was without significant effect. Optogenetic stimulation of projection neurons from the substantia nigra (SN) to the DMS, aligned either with reward receipt or omission or prior to making a choice, had no significant effect on reversal learning. These findings are consistent with the notion that increased activity in the VTA → NAcS pathway disrupts behavioural flexibility by impairing learning from negative reinforcement.

Federated Learning for Predicting Postoperative Remission of Patients with Acromegaly: A Multicentered study.

Decentralized federated learning (DFL) may serve as a useful framework for machine learning (ML) tasks in multicentered studies, maximizing the use of clinical data without data sharing. We aim to propose the first workflow of DFL for ML tasks in multicentered studies, which can be as powerful as those using centralized data. A DFL workflow was developed with four steps: registration, local computation, model update, and inspection. A total of 598 participants with acromegaly from PUMCH, and 120 participants from XWH were enrolled. The cohort from PUMCH was further split into five centers. Nine clinical features were incorporated into ML-based models trained based on four algorithms: LR, GBDT, SVM, and DNN. The area under the curve (AUC) of receiver operating characteristic curves was used to evaluate the performance of the models. Models trained based on DFL workflow performed better than most models in LR (P<0.05), all models in DNN, SVM, and GBDT (P<0.05). Models trained on DFL workflow performed as powerful as models trained on centralized data in LR, DNN, and SVM (P>0.05). We demonstrate that the DFL workflow without data sharing should be a more appropriate method in ML tasks in multicentered studies. And the DFL workflow should be further exploited in clinical researches in other departments and it can encourage and facilitate multicentered studies.

Diffraction‐Driven Parallel Convolution Processing with Integrated Photonics

AbstractTraditional electronic processors often struggle with bandwidth limitations and high power consumption when executing extensive linear operations for deep learning tasks. Optical computing has emerged as a promising alternative, offering parallel and energy‐efficient computation capabilities. Yet, the development of high‐density optical computing architectures on integrated photonic platforms remains limited, hindered by constraints in neuron scalability and control engineering complexities. Addressing these challenges, this work presents a diffraction‐driven multi‐kernel optical convolution unit (MOCU) that enables on‐chip parallel convolution processing. By utilizing cascaded silica 1D metalines as pre‐trained large‐scale weights and employing spatial multiplexing at the output, MOCU allows simultaneous passive computation of diverse convolutions within a single unit. This architecture facilitates the construction of optical convolutional neural networks (OCNNs), enabling efficient machine vision processing with a streamlined design. To mitigate errors in MOCU‐embedded OCNNs, a lightweight electronic neural network operates concurrently to calibrate systematic deviations via a low‐rank adaptation (LoRA) algorithm, with minimal overhead. The fabricated MOCU chip demonstrates the highest independent 8‐kernel convolutions in parallel, each with a kernel size and occupying just 0.06 . This architecture effectively merges photonic and electronic technologies, offering a scalable design pathway for energy‐efficient, high‐density deep learning hardware.

A framework for measuring the training efficiency of a neural architecture

Measuring Efficiency in neural network system development is an open research problem. This paper presents an experimental framework to measure the training efficiency of a neural architecture. To demonstrate our approach, we analyze the training efficiency of Convolutional Neural Networks and Bayesian equivalents on the MNIST and CIFAR-10 tasks. Our results show that training efficiency decays as training progresses and varies across different stopping criteria for a given neural model and learning task. We also find a non-linear relationship between training stopping criteria, training Efficiency, model size, and training Efficiency. Furthermore, we illustrate the potential confounding effects of overtraining on measuring the training efficiency of a neural architecture. Regarding relative training efficiency across different architectures, our results indicate that CNNs are more efficient than BCNNs on both datasets. More generally, as a learning task becomes more complex, the relative difference in training efficiency between different architectures becomes more pronounced.

Interpreting pretext tasks for active learning: a reinforcement learning approach

As the amount of labeled data increases, the performance of deep neural networks tends to improve. However, annotating a large volume of data can be expensive. Active learning addresses this challenge by selectively annotating unlabeled data. There have been recent attempts to incorporate self-supervised learning into active learning, but there are issues in utilizing the results of self-supervised learning, i.e., it is uncertain how these should be interpreted in the context of active learning. To address this issue, we propose a multi-armed bandit approach to handle the information provided by self-supervised learning in active learning. Furthermore, we devise a data sampling process so that reinforcement learning can be effectively performed. We evaluate the proposed method on various image classification benchmarks, including CIFAR-10, CIFAR-100, Caltech-101, SVHN, and ImageNet, where the proposed method significantly improves previous approaches.

Mitigating Catastrophic Forgetting in Continual Learning for Natural Language Processing Tasks

Mitigating Catastrophic Forgetting in Continual Learning for Natural Language Processing Tasks

TVGeAN: Tensor Visibility Graph-Enhanced Attention Network for Versatile Multivariant Time Series Learning Tasks

This paper introduces Tensor Visibility Graph-enhanced Attention Networks (TVGeAN), a novel graph autoencoder model specifically designed for MTS learning tasks. The underlying approach of TVGeAN is to combine the power of complex networks in representing time series as graphs with the strengths of Graph Neural Networks (GNNs) in learning from graph data. TVGeAN consists of two new main components: TVG which extend the capabilities of visibility graph algorithms in representing MTSs by converting them into weighted temporal graphs where both the nodes and the edges are tensors. Each node in the TVG represents the MTS observations at a particular time, while the weights of the edges are defined based on the visibility angle algorithm. The second main component of the proposed model is GeAN, a novel graph attention mechanism developed to seamlessly integrate the temporal interactions represented in the nodes and edges of the graphs into the core learning process. GeAN achieves this by using the outer product to quantify the pairwise interactions of nodes and edges at a fine-grained level and a bilinear model to effectively distil the knowledge interwoven in these representations. From an architectural point of view, TVGeAN builds on the autoencoder approach complemented by sparse and variational learning units. The sparse learning unit is used to promote inductive learning in TVGeAN, and the variational learning unit is used to endow TVGeAN with generative capabilities. The performance of the TVGeAN model is extensively evaluated against four widely cited MTS benchmarks for both supervised and unsupervised learning tasks. The results of these evaluations show the high performance of TVGeAN for various MTS learning tasks. In particular, TVGeAN can achieve an average root mean square error of 6.8 for the C-MPASS dataset (i.e., regression learning tasks) and a precision close to one for the SMD, MSL, and SMAP datasets (i.e., anomaly detection learning tasks), which are better results than most published works.

A novel image classification framework based on variational quantum algorithms

Image classification is a crucial task in machine learning with widespread practical applications. The existing classical framework for image classification typically utilizes a global pooling operation at the end of the network to reduce computational complexity and mitigate overfitting. However, this operation often results in a significant loss of information, which can affect the performance of classification models. To overcome this limitation, we introduce a novel image classification framework that leverages variational quantum algorithms (VQAs) hybrid approaches combining quantum and classical computing paradigms within quantum machine learning. The major advantage of our framework is the elimination of the need for the global pooling operation at the end of the network. In this way, our approach preserves more discriminative features and fine-grained details in the images, which enhances classification performance. Additionally, employing VQAs enables our framework to have fewer parameters than the classical framework, even in the absence of global pooling, which makes it more advantageous in preventing overfitting. We apply our method to different state-of-the-art image classification models and demonstrate the superiority of the proposed quantum architecture over its classical counterpart through a series of state vector simulation experiments on public datasets. Our experiments show that the proposed quantum framework achieves up to a 9.21% increase in accuracy and up to a 15.79% improvement in F1 score, compared to the classical framework. Additionally, we explore the impact of shot noise on our method through shot-based simulation and find that increasing the number of measurements does not always lead to better results. Selecting an appropriate number of measurements can yield optimal results, even surpassing those obtained from state vector simulation.

Human-AI Collaboration: A Student-Centered Perspective of Generative AI Use in Higher Education

The increasing adoption of generative AI (Gen AI) has made it even more important to investigate how education and learning is transformed. The paper investigates the evolving relationship between humans and AI in performing learning and knowledge-intensive tasks. Based on mixed data collected from business school students, the article explores the evolving relationships between human and AI in knowledge collaboration. The article sets out to address how Gen AI usage affects students’ behavior, their academic work and their attitudes towards AI. It investigates how students use Gen AI in an academic context to support students’ work processes. Our analysis draws on collaborative learning theories and revisits the automation-augmentation paradox in the context of management education. The aim is to address the following research questions: How does collaboration with AI affect student academic work? And how do students in higher education use Gen AI for academic work? We present insights into students’ perception of benefits and drawbacks of using Gen AI, their attitude towards Gen AI, knowledge tasks in which they collaborate or delegate to Gen AI and we discuss the potential risks associated with the use of Gen AI. Students are collaborating with AI using ChatGPT for expanding knowledge on academic themes, summarizing concepts, theories, and generating ideas for research topics and methods. Time saving, enhanced productivity and user-friendliness of the tool were identified as the main benefits associated with Gen AI, whereas the risk of plagiarism, its inaccuracy in responses and the need for prior knowledge were identified as the main drawbacks. While our findings underscore the potential of Gen AI to significantly enhance student learning experiences, they also underscore the importance of exercising caution and awareness of associated risks to automate learning. This study seeks to enrich our comprehension of AI's transformative role in higher education, with a specific focus on the student-centered perspective.

Graph neural processes for molecules: an evaluation on docking scores and strategies to improve generalization

Neural processes (NPs) are models for meta-learning which output uncertainty estimates. So far, most studies of NPs have focused on low-dimensional datasets of highly-correlated tasks. While these homogeneous datasets are useful for benchmarking, they may not be representative of realistic transfer learning. In particular, applications in scientific research may prove especially challenging due to the potential novelty of meta-testing tasks. Molecular property prediction is one such research area that is characterized by sparse datasets of many functions on a shared molecular space. In this paper, we study the application of graph NPs to molecular property prediction with DOCKSTRING, a diverse dataset of docking scores. Graph NPs show competitive performance in few-shot learning tasks relative to supervised learning baselines common in chemoinformatics, as well as alternative techniques for transfer learning and meta-learning. In order to increase meta-generalization to divergent test functions, we propose fine-tuning strategies that adapt the parameters of NPs. We find that adaptation can substantially increase NPs' regression performance while maintaining good calibration of uncertainty estimates. Finally, we present a Bayesian optimization experiment which showcases the potential advantages of NPs over Gaussian processes in iterative screening. Overall, our results suggest that NPs on molecular graphs hold great potential for molecular property prediction in the low-data setting.Scientific contributionNeural processes are a family of meta-learning algorithms which deal with data scarcity by transferring information across tasks and making probabilistic predictions. We evaluate their performance on regression and optimization molecular tasks using docking scores, finding them to outperform classical single-task and transfer-learning models. We examine the issue of generalization to divergent test tasks, which is a general concern of meta-learning algorithms in science, and propose strategies to alleviate it.

The Students Satisfaction and Interest in Roller Skating

This study is based on classroom satisfaction and students' roller skating course learning interest, laying a foundation for improving students' roller skating course learning quality. The results show that students' learning satisfaction and roller skating course learning interest in classroom learning have been improved, which has improved the learning quality of roller skating courses. Through descriptive correlation design and SPSS analysis of data from 300 students, it was found that students' roller skating interests in high-level dimensions, including self-flexibility; Roller skating skills and fundamentals; personal interest; Parents; Teachers; Facilities; Roller skating course and roller skating course satisfaction Learning content; Learning resources and activities; Learning task Implementation, etc. In order to improve students' classroom satisfaction and roller skating course learning interest, teachers should take relevant measures in teaching, such as not considering sex, age, as the main factors for evaluating students' learning interest and satisfaction, emphasizing the actual content of course design and teachers' organizational means, helping students build interest, and strengthening the development quality of roller skating courses.