Learning Performance Research Articles

Deep Feature Retention Module Network for Texture Classification

Texture describes the unique features of an image. Therefore, texture classification is a crucial task in computer vision. Various CNN-based deep learning methods have been developed to classify textures. During training, the deep-learning model undergoes an end-to-end procedure of learning features from low to high levels. Most CNN architectures depend on high-level features for the final classification. Hence, other low- and mid-level information was not prioritized for the final classification. However, in the case of texture classification, it is essential to determine detailed feature information within the pattern to classify textures as they have diversity and irregularity in images within the same class. Therefore, the feature information at the low- and mid-levels can also provide meaningful information to distinguish the classes. In this study, we introduce a CNN model with a feature retention module (FRM) to utilize features from numerous levels. FRM maintains the texture information extracted at each level and extracts feature information through filters of various sizes. We used three texture datasets to evaluate the proposed model combined with the FRM. The experimental results showed that learning using different levels of features together assists in improving learning performance more than learning using high-level features.

Empowering learners with AI‐generated content for programming learning and computational thinking: The lens of extended effective use theory

AbstractBackgroundArtificial intelligence–generated content (AIGC) has stepped into the spotlight with the emergence of ChatGPT, making effective use of AIGC for education a hot topic.ObjectivesThis study seeks to explore the effectiveness of integrating AIGC into programming learning through debugging. First, the study presents three levels of AIGC integration based on varying levels of abstraction. Then, drawing on extended effective use theory, the study proposes the underlying mechanism of how AIGC integration impacts programming learning performance and computational thinking.MethodsThree debugging interfaces integrated with AIGC by ChatGPT were developed for this study according to three levels of AIGC integration design. The study conducts a between‐subject experiment with one control group and three experimental groups. Analysis of covariance and a structural equation model are employed to examine the effects.Results and ConclusionsThe results show that the second and third levels of abstraction in AIGC integration yield better learning performance and computational thinking, but the first level shows no difference compared to traditional debugging. The underlying mechanism suggests that the second and third levels of abstraction promote transparent interaction, which enhances representational fidelity and consequently impacts learning performance and computational thinking, as evidenced in test of the mechanism. Moreover, the study finds that learning fidelity weakens the effect of transparent interaction on representational fidelity. Our research offers valuable theoretical and practical insights.

The Effect of University Students’ Academic Self-Efficacy on Academic Burnout: The Chain Mediating Role of Intrinsic Motivation and Learning Engagement

Academic burnout seriously affects the academic performance and mental health of college students. This study developed a multi-mediation model to investigate the relationship between academic self-efficacy and academic burnout. A total of 1431 undergraduate students (51.85% female) were recruited to participate in this study. And we used SPSS PROCESS developed by Hayes to examine the mediating role of academic self-efficacy and academic burnout. The results indicated that academic self-efficacy can negatively predict academic burnout. Moreover, intrinsic motivation and learning engagement have a significant mediating effect between academic self-efficacy and academic burnout. Finally, we constructed a model comprising four variables and found that it explained 47.62% of the variance in academic burnout. These findings can be conducive to a suitable intervention path to reduce academic burnout of Chinese college students, so as to develop effective instructional strategies to improve their learning performance.

Absolute Value Inequality SVM for the PU Learning Problem

Positive and unlabeled learning (PU learning) is a significant binary classification task in machine learning; it focuses on training accurate classifiers using positive data and unlabeled data. Most of the works in this area are based on a two-step strategy: the first step is to identify reliable negative examples from unlabeled examples, and the second step is to construct the classifiers based on the positive examples and the identified reliable negative examples using supervised learning methods. However, these methods always underutilize the remaining unlabeled data, which limits the performance of PU learning. Furthermore, many methods require the iterative solution of the formulated quadratic programming problems to obtain the final classifier, resulting in a large computational cost. In this paper, we propose a new method called the absolute value inequality support vector machine, which applies the concept of eccentricity to select reliable negative examples from unlabeled data and then constructs a classifier based on the positive examples, the selected negative examples, and the remaining unlabeled data. In addition, we apply a hyperparameter optimization technique to automatically search and select the optimal parameter values in the proposed algorithm. Numerical experimental results on ten real-world datasets demonstrate that our method is better than the other three benchmark algorithms.

Modal-aware prompt tuning with deep adaptive feature enhancement

Prompt learning has recently emerged as a promising method for fine-tuning vision-language models. By introducing prompts in the text encoder or image encoder, the pre-trained model can quickly adapt to downstream tasks without updating the pre-trained weights. However, prior multi-modal prompt tuning works do not consider the difference in feature distributions between text and images, and adopt the same prompts for both encoders, thus achieving sub-optimal performance in the downstream few-shot learning. In this paper, we propose Modal-Aware Prompt (MAP) to alleviate this issue. Specifically, considering the stability of text features, we design text-specific prompts, which can acquire text class-related information from a general template (i.e., “a photo of a <category>”) by unidirectional attention-based interaction. Additionally, considering the diversity of image features, we design visual-specific prompts to acquire image class-related information and adjust the image features by bidirectional attention-based interaction. To learn hierarchical prompt representations and reinforce the prompt features, we further propose a Deep Adaptive Feature Enhancement (DAFE) module to adaptively utilize the prompt output of the former layer, which can combine instance-level and task-level information simultaneously. Combining the above two designs, our method MAP-DAFE obtains state-of-the-art results on 11 image recognition datasets and has the fastest convergence rate. This proves our MAP-DAFE is effective and efficient.

An Analysis of Student's Academic Achievement in Genetic Algorithms Based on Mobile Learning Applications

Several scholars have focused on this area due to the significance of mobile technology in the educational process, resulting in a substantial body of scholarly work. The primary objective of this study is to investigate the impact of mobile learning on students’ academic performance. The meta-analysis approach was utilized in this research. The available research was examined using several databases to find the relevant studies that were within the scope of the investigation. The study’s inclusion criteria and components were implemented following a literature review. This research proposes a mobile education system approach based on genetic algorithms to address the issues with the current system. The analysis findings indicate that most of them stay at around 10%. The salary only slightly increases when numerous individuals teach, but it also falls within a regular, appropriate range that can accommodate more complex tasks. According to the factor analysis findings, the influence of mobile educational devices on students’ learning performance varied depending on the course and subject. However, it remained constant regardless of the students’ education level and implementation duration. Apart from the previously mentioned research findings, this article also includes a descriptive analysis of the studies that were part of the meta-analysis.

Improved financial forecasting via quantum machine learning

Quantum algorithms have the potential to enhance machine learning across a variety of domains and applications. In this work, we show how quantum machine learning can be used to improve financial forecasting. First, we use classical and quantum Determinantal Point Processes to enhance Random Forest models for churn prediction, improving precision by almost 6%. Second, we design quantum neural network architectures with orthogonal and compound layers for credit risk assessment, which match classical performance with significantly fewer parameters. Our results demonstrate that leveraging quantum ideas can effectively enhance the performance of machine learning, both today as quantum-inspired classical ML solutions, and even more in the future, with the advent of better quantum hardware.

Integrating Artificial Intelligence (AI) Into Adult Education

This conceptual article provides a comprehensive overview of the current status of Artificial Intelligence (AI) integration and its influence on adult education. It discusses generative AI technologies and their potential applications in adult education settings, examines the opportunities and ethical challenges associated with integrating AI, and provides insights into emerging trends. The article consists of five sections. The introduction provides a rationale as to why AI should be integrated into adult education. Second, it describes evolving AI technologies such as Large Language Models (LLM) for personalized learning, Machine Learning Algorithms for adaptive learning systems, Virtual Reality (VR) and Augmented Reality (AR) for immersive learning experiences, Chatbots and virtual assistants for learner support and guidance, and Data Learning Analytics (DLA) for tracking learner progress and performance into adult education. Section three explores the ethical implications of AI in adult education, including academic honesty and integrity, data privacy, and algorithmic bias. In section four, emerging trends and future directions are discussed. The final section considers policy implications and makes recommendations for adult educators working to develop AI-enriched adult education.

Distributed statistical learning algorithm for nonlinear regression with autoregressive errors

The growing size of modern data brings challenges to statistical learning, and substantial distributed algorithms have been proposed. However, most of them need the homogeneity assumption that the distribution of the local data is the same as that of the global data. This is seldom in practice, and the learning performance deteriorates seriously if this assumption is not satisfied. Moreover, they are only for independent data, and cannot incorporate the serial correlations between data. To solve these issues, we propose a novel distributed statistical learning framework for the nonlinear regression with autoregressive errors, which realizes communication-efficient distributed optimization, and overcomes the homogeneity assumption. The theoretical results also guarantee that the new distributed framework is equivalent to the global one. Numerical experiments also illustrate the good performance of the new method.

Promoting students' creative self-efficacy, critical thinking and learning performance: An online interactive peer assessment approach guided by constructivist theory in maker activities

Peer assessment is acknowledged as a potent strategy that motivates students to engage in reflection and comparison through the evaluation of their peers' work. However, the effectiveness of peer assessment is not always satisfactory due to a number of factors, such as lack of in-depth understanding of peers' work and insufficient motivation for the behavior of evaluation. Based on constructivist theory, this study proposed an Understanding-Evaluation-Backward Evaluation-Reflection based peer assessment (UEBR-PA) approach, which directs students to comprehensively grasp their peers' projects, undertake efficient evaluations, provide retrospective reviews on others' feedback, and ultimately, integrate this collective feedback to refine and enhance their own projects. To evaluate the effectiveness of the proposed approach, a quasi-experiment was conducted in an Open-Source Hardware Project Design course with 54 university students from two classes. One class was assigned as the experimental group to conduct the peer assessment using the proposed approach, while the other class was assigned as the control group to conduct the peer assessment using the conventional presentation-evaluation-communication based approach. The results of the study indicated that the UEBR-PA approach significantly increased students' creative self-efficacy, critical thinking tendency, and learning performance. Furthermore, students engaging with the UEBR-PA approach showcased more positive interactive assessment behaviors.

Do attitudes matter? An investigation into students’ attitudes toward engineering and learning performances through engineering-focused STEM curricula

Do attitudes matter? An investigation into students’ attitudes toward engineering and learning performances through engineering-focused STEM curricula

The dynamics of functional brain network segregation in feedback-driven learning

Prior evidence suggests that increasingly efficient task performance in human learning is associated with large scale brain network dynamics. However, the specific nature of this general relationship has remained unclear. Here, we characterize performance improvement during feedback-driven stimulus-response (S-R) learning by learning rate as well as S-R habit strength and test whether and how these two behavioral measures are associated with a functional brain state transition from a more integrated to a more segregated brain state across learning. Capitalizing on two separate fMRI studies using similar but not identical experimental designs, we demonstrate for both studies that a higher learning rate is associated with a more rapid brain network segregation. By contrast, S-R habit strength is not reliably related to changes in brain network segregation. Overall, our current study results highlight the utility of dynamic functional brain state analysis. From a broader perspective taking into account previous study results, our findings align with a framework that conceptualizes brain network segregation as a general feature of processing efficiency not only in feedback-driven learning as in the present study but also in other types of learning and in other task domains.

Early Prediction of Students Performance in Higher Education

Evaluating students' learning performance is a fundamental aspect of evaluating any educational institution. When addressing challenges related to the learning process, student performance is critical, and it is one of the key factors used to quantify learning outcomes. The topic of research known as educational data mining (EDM) has grown out of the potential to leverage data knowledge to enhance educational systems. EDM is the development of methods to analyze data collected from educational environments, enabling a more complete and precise understanding of students and the enhancement of their educational results. Evaluating the students' learning results is a crucial part of evaluating any educational institution. One of the key variables used to quantify learning outcomes is student performance, which is significant when addressing problems with the learning process. The field of research known as educational data mining, or EDM, was born out of the potential to leverage data knowledge to enhance educational systems. EDM is the process of developing methods for evaluating information obtained from educational environments. This makes it possible to learn more precise and in-depth information about students and enhances their academic achievement. Academic achievement tests (AAT), general aptitude tests (GAT), admission scores, first-level courses, and other early-stage factors are used in the paper's dimensionality reduction mechanism by T-SNE algorithm for the clustering technique. This allows the study to investigate the relationship between these aspects and GPAs. Regarding the categorization method, the study showcases tests conducted on various machine learning models that forecast student achievement in the initial phases by utilizing diverse attributes such as course grades and entrance exam results. To gauge the models' quality, we employ various evaluation measures. Based on the findings, it appears that early student failure rates can be reduced by educational institutions.

Effectiveness of GeoGebra on Botswana TVET students’ performance on linear equations

The purpose of this study was to investigate the comparative effects of using GeoGebra software and traditional teaching approaches on the performance of TVET students in linear equations. The study followed a quasi-experimental design with a non-equivalent group consisting of pre- and post-test measures. The population of the study consisted of TVET learners from 14 Brigades in the southern region of Botswana. Using the convenience sampling technique, two TVET institutions (n=41) participated in the study. One institution formed the experimental group, whereas the other formed the control group. The study did not employ random participant selection, as the control group and the experimental group comprised pre-existing, intact classes from the participating institutions. GeoGebra software was used in the experimental group, while the control group was taught using the traditional approach with no technology incorporation. Similar pre- and post-tests were administered to both groups to measure the comparative effects of each teaching method on the performance of learners. A constructivist theory framed the study. A t-test was used to analyze the data with α = 0.05, and the findings showed a significant difference in the mean scores between the control group and the experimental group (t = 2.188, p = 0.008 &lt; α) in the post-test. These results indicated that students in the experimental group outperformed those in the control group. Overall, the results of this study showed that GeoGebra enhanced student performance in linear equations as compared to traditional methods of teaching.

Assessment of soil salinity using explainable machine learning methods and Landsat 8 images

The aim of this study is to comparatively analyze the performance of machine learning (ML) algorithms for modeling soil salinity using field-based electrical conductivity (EC) data and Landsat-8 OLI satellite images with derived environmental covariates. We also aim to interpret and explain the ML models with and without over-sampling methods using Shapley (SHAP) values, an explainable ML approach that has not yet been utilized for soil salinity estimation tasks as an ML problem. We investigate two case study areas from western and southeastern Lake Urmia Playas (LUP) in the Northwest of Iran. Our study uses 26 environmental covariates, two ML models, namely extreme gradient boosting (XGBoost) and random forest (RF), and two over-sampling methods: synthetic minority over-sampling technique (SMOTE) and random over-sampling (ROS). Results indicate that XGBoost performs better compared to RF in terms of both R2 and RMSE. Additionally, the visual interpretation of soil salinity maps demonstrated the superiority of XGBoost. SMOTE produced superior results than ROS and no over-sampling test cases. Finally, SHAP analysis illustrated that vegetation indices made a greater contribution to the soil salinity prediction in the West LUP, while visible bands contributed more in the Southeast LUP Region.

Meta-learning approaches for few-shot learning: A survey of recent advances

Despite its astounding success in learning deeper multi-dimensional data, the performance of deep learning declines on new unseen tasks mainly due to its focus on same-distribution prediction. Moreover, deep learning is notorious for poor generalization from few samples. Meta-learning is a promising approach that addresses these issues by adapting to new tasks with few-shot datasets. This survey first briefly introduces meta-learning and then investigates state-of-the-art meta-learning methods and recent advances in: (i) metric-based, (ii) memory-based, (iii), and learning-based methods. Finally, current challenges and insights for future researches are discussed.

The Pathway to Developing a Competency-based Tool for Entry-level Nursing Simulation: A Delphi Study

BackgroundSimulation-based learning experiences allow undergraduate nursing students to develop competence and confidence through deliberate practice with immediate feedback on the learner's performance through debriefing. With the transition to competency-based nursing education, nursing faculty need more guidance in implementing competency-based evaluations in the simulation setting. PurposeThis Delphi study aims to inform the future development of a competency-based tool – SimComp – based on the American Association of Colleges of Nursing (AACN) Essentials. MethodsA Delphi framework was used to recruit expert nursing faculty to complete the surveys via an online platform. Data analysis occurred through open-ended questions and quantitative methods to ensure that the responses from expert panelists were used to form the results. ResultsAfter four rounds of this Delphi study, a consensus was achieved on 111 appropriate items for assessing competence in the simulation-based learning environment. ConclusionWhile further research is warranted, this study provides insight for nursing institutions considering implementing or increasing the use of simulation within their program for competency-based evaluations.

Abstract PO2-07-05: Deep learning model for automated quantification of HER2 expression in invasive breast cancers from immunohistochemical whole slide images

Abstract Introduction Human epidermal growth factor receptor 2 (HER2) protein overexpression and/or HER2 gene amplification is found in about 20% of invasive breast cancers. Considering the results of the DESTINY-Breast trials confirming the remarkable efficacy of anti-HER2 antibody drug-conjugate (T-Dxd) in both HER2-overexpressed and HER2-low tumors, it is necessary to identify not only HER2 (immunohistochemistry (IHC) score 3+) overexpressing tumors, but also HER2-low tumors. The latter category, defined as IHC1+ or IHC2+ but non-amplified, has proven to be challenging even for experienced pathologists, with high inter-observer variability. Here, we validate the performance of a deep learning (DL) model at: 1) predicting the IHC score from IHC histological features and 2) identifying HER2-low tumors with high sensitivity. Methods 675 HER2 stained IHC slides from primary breast cancer patients were selected based on pathology reports across three different cohorts (KCL_GSTT_1 n=369; Cypath_Breast n=214; KCL_GSTT_2 n=92). The slides were digitized and reviewed by expert breast pathologists. Specifically, Cypath_Breast and KCL_GSTT_2 were each annotated by one expert pathologist, while KCL_GSTT_1 was annotated by 5 expert pathologists with the ground truth defined as a majority vote between the 5 annotators. Cypath_Breast was assigned as the “discovery” set; while KCL_GSTT_1 and KCL_GSTT_2 were used as two independent validation cohorts. The model was specifically trained to extract features from IHC tiles to ensure they would capture both the staining intensity and the staining location on the cells (membrane, cytoplasmic or nuclei). The one-vs-one (OVO) and one-vs-rest (OVR) AUC as well as sensitivity and specificity to HER2-low and positive tumors were used as metrics to assess the performance of our model. An expert pathologist reviewed the most predictive regions of HER2 expression. Results After review by expert pathologists, 59 slides from KCL_GSTT_1 were removed either because of folded tissue, blurriness or because there was too little tissue. The IHC scores (0/1+/2+/3+) for the 3 cohorts were distributed as follows; Cypath_Breast: 42/120/36/16; KCL_GSTT_1: 120/90/52/48; KCL_GSTT_2: 54/22/15/1. For KCL_GSTT_1, the average agreement between the 5 expert pathologists amounted to a Cohen’s Kappa score of 0.63. Following training on Cypath_Breast, the performance of the model are presented in An ablation study proved that a feature extractor trained on IHC tiles outperformed an in-house extractor solely trained on H&amp;E tiles (OVO AUC 0.95 vs 0.93 on KCL_GSTT_1; DeLong p&amp;lt; 0.001). Specifically, the model performed best at distinguishing HER2-negative (HER2 IHC 0) from HER2-low and HER2-positive tumors with a sensitivity (0 vs 1+/2+/3+) of 0.95 [0.93 - 0.98] and a specificity (0 vs 1+/2+/3+) of 0.78 [0.72 - 0.84] on KCL_GSTT_1. We observed the same tendency on KCL_GSTT_2 with a sensitivity (0 vs 1+/2+/3+) of 0.97 [0.91 - 1.00] and a specificity of 0.91 [0.83 - 0.98]. More importantly, the DL model reached a Cohen’s Kappa score of 0.72 [0.67 - 0.77] with the ground truth on KCL_GSTT_1, surpassing the average agreement Kappa score between expert pathologists (0.63). The most predictive regions showed cytoplasmic staining in tumor regions. Conclusion Our model provides a path towards a fully automated workflow for identifying up to 95% of breast cancer patients who could potentially benefit from anti-HER2 targeted therapies. Additionally, we show the utility of AI-based tools to minimize discrepancies among pathologists and assist in the diagnostic patient pathway. Table 1: Results of external validation on KCL_GSTT_1 and KCL_GSTT_2 &amp;lt;grpahic&amp;gt; Cohorts OVO AUC OVR AUC IHC 0 OVR AUC IHC 1+ OVR AUC IHC 2+ OVR AUC IHC 3+ KCL_GSTT_1 0.95 0.95 0.80 0.91 0.99 [0.93 - 0.96] [0.94 - 0.95] [0.79 - 0.82] [0.90 - 0.92] [0.99 - 1.00] KCL_GSTT_2 0.92 0.97 0.78 0.84 N/A [0.90 - 0.93] [0.96 - 0.98] [0.76 - 0.81] [0.80 - 0.88] Citation Format: Pierre-Antoine Bannier, Loïc Herpin, Rémy Dubois, Lydwine Van Praet, Charles Maussion, Ellen Amonoo, Anca Mera, Jasmine Timbres, Cheryl Gillett, Elinor Sawyer, Patrycja Gazinska, Piotr Ziolkowski, Roberto Salgado, Sheeba Irshad. Deep learning model for automated quantification of HER2 expression in invasive breast cancers from immunohistochemical whole slide images [abstract]. In: Proceedings of the 2023 San Antonio Breast Cancer Symposium; 2023 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2024;84(9 Suppl):Abstract nr PO2-07-05.

Investigation of Mild Traumatic Brain Injury Home Cage Behavior: The Home Cage Assay Advantages.

This study utilized the Noldus PhenoTyper Home Cage Monitoring system (HCM) to assess the behavioral and cognitive changes of experimental closed-head mild traumatic brain injury (mTBI). Seventy-nine adult male Institute of Cancer Research (ICR) mice were subjected to either a sham procedure or closed-head mTBI using the weight-drop model. Seven days post-injury, separate cohorts of mice underwent either a non-cognitive or a cognitive home cage assessment, a treadmill fatigue test, or the Open Field Test. mTBI significantly influenced habituation behavior and circadian wheel-running activity. Notably, mTBI mice exhibited an increased frequency of visits to the running wheel, but each visit was shorter than those of controls. No significant differences between the groups in discrimination or reversal learning performance were observed. However, during the reversal learning stage, mTBI mice performed similarly to their initial discrimination learning levels, suggesting an abnormally faster rate of reversal learning. Home cage monitoring is a valuable tool for studying the subtle effects of mTBI, complementing traditional assays. The automated evaluation of habituation to novel stimuli (e.g., novel environment) could serve as a potentially sensitive tool for assessing mTBI-associated behavioral deficits.

Diagnostic performance of machine learning in systemic infection following percutaneous nephrolithotomy and identification of associated risk factors

ObjectiveThis study aims to investigate the predictive performance of machine learning in predicting the occurrence of systemic inflammatory response syndrome (SIRS) and urosepsis after percutaneous nephrolithotomy (PCNL). MethodsA retrospective analysis was conducted on patients who underwent PCNL treatment between January 2016 and July 2022. Machine learning techniques were employed to establish and select the best predictive model for postoperative systemic infection. The feasibility of using relevant risk factors as predictive markers was explored through interpretability with Machine Learning. ResultsA total of 1067 PCNL patients were included in this study, with 111 (10.4 %) patients developing SIRS and 49 (4.5 %) patients developing urosepsis. In the validation set, the risk model based on the GBM protocol demonstrated a predictive power of 0.871 for SIRS and 0.854 for urosepsis. Preoperative and postoperative platelet changes were identified as the most significant predictors. Both thrombocytopenia and thrombocytosis were found to be risk factors for SIRS or urosepsis after PCNL. Furthermore, it was observed that when the change in platelet count before and after PCNL surgery exceeded 30*109/L (whether an increase or decrease), the risk of developing SIRS or urosepsis significantly increased. ConclusionMachine learning can be effectively utilized for predicting the occurrence of SIRS or urosepsis after PCNL. The changes in platelet count before and after PCNL surgery serve as important predictors.