Effective Learning Research Articles

대학 교양수업에서 학습자 중심의 팀기반 학습 경험

Objectives The purposes of this study were to explores the experiences that learners who participate in team-based learning experience during the learning process, examines the learners' experiences in team-based learning in a contextual setting through a qualitative approach, and seeks educational reflection and implications. Methods For this purpose, this study was conducted on 94 college students who participated in liberal arts classes. To achieve the purpose of the study, first, team-based learning was designed by selecting a ‘pedagogy-based lib-eral arts’ subject. Second, after analyzing the content covered in the selected subject, the instructor's lecture on the learning topic was conducted in advance in the first half of each lesson, and students studied individually through lectures on the learning topic. After watching various learning materials related to the learning topic, team-based learning related to the learning content was applied. Third, we analyzed individual and team activity sheets, study notes for each session, and reflection logs written by learners before and after learning. The response sheets used in the analysis were all qualitative data collected through classes, and were analyzed through the inter-action of categorization, coding, structuring, and interpretation during the data analysis process. Results In terms of learning motivation from the results of this study, first, the motivation for taking the course was participation due to obligation, accidental choice, and unclear expectations due to the nature of liberal arts subjects, and second, the topic and content of learning were based on their own real-life usefulness and career-related topics. Third, actively apply the cognitive information acquired through participation in learning to solve life problems. Fourth, setting learning goals shows educationally important information and context for converting extrinsic motivation into intrinsic motivation. In terms of cognition, learners showed the highest cognitive factors in motivation to take the course, personal learning goals, and learner needs. In terms of interaction, in the process of participating in team-based learning, learners placed first, the attitude of participants, including communication, participation, and courtesy among team members, as the most important, and second, they meaningfully experienced positive experi-ences and emotional support. In terms of teaching and learning, positive experiences such as self-expression, dis-cussion and presentation, initiative, and overcoming presentation anxiety were recognized through participation in team learning. On the other hand, negative aspects could act as a hindrance to positive interactions between team members. In particular, to overcome these obstacles in the early stages of learning, multifaceted efforts were need-ed, such as clear motivation, setting learning goals, setting learning content that suits the learner's needs and inter-ests, emotional support from the instructor, effective learning support materials, and evaluation plans. Conclusions Through this study, it can be seen that, due to the nature of the learning process in team-based classes where learners' interaction is the center, more active and interactive communication and voluntary and proactive participation are required than previous learning experiences.To achieve this, active and flexible re-sponses will be needed to improve the anxiety and dissatisfaction that learners may feel. Through this, the charac-teristics and diversity of individual learners can be mutually recognized and accepted at the beginning of learning, the strengths of team-based learning of emotional support and consideration through interaction between team members can be effectively demonstrated, and learners' expectations and demands for participation in learning can be improved.

Pedagogical relations as a decolonisation tool in African higher education: Reflection on the ethics of care, respect, and trust

This article uses the findings obtained from a study that delves into the perceptions of students on their relationship with their higher education (HE) lecturers and how it affects their academic success, to respond to the issue of decolonisation in South African HE, and to approach the question of decolonising HE in Africa. The article argues that it is essential to prioritise student well-being, amplify their voices, and promote a caring culture towards addressing the issue of decolonisation in African education systems. The study shows that African HE students hold higher expectations of their lecturers beyond being professionals. This expectation includes respect for students’ thoughts and incorporation of epistemology that aligns with fostering African development, culture and thoughts (without necessarily conforming to neoliberal norms). The four categories of the teacher’s role, which include academic development, respect and trust, social relationships, and ethics of care are highly demanded by HE students. Borrowing the study outcome, this article holds the view that students’ high expectations of their lecturers to foster social relationships should be channelled to incorporate the African student as a collaborator in the business of education and as a response to the demands of HE students to decolonise African education system. These four categories are not only institutional strategies for effective teaching and learning but also a way to address the non-inclusive impact of Western epistemology on historically racial institutions in Africa.Contribution: This article proposes that adopting mainstream pedagogical relationships can be a powerful tool in incorporating the African students’ thoughts and a step towards liberating the HE system in Africa. It recommends these four cardinal themes as institutional strategies for restricting teaching and learning that relegate students to the receiving end thus systems that refute students as collaborators of knowledge sharing especially at historically racial institutions.

Bibliometric Analysis: Research Trends in Creative Thinking Behavior in Learning

The ability to think outside the box, create something that has never existed before, and apply knowledge and skills creatively and inventively. Effective learning activities require a balance between several aspects, namely cognitive, affective, and psychomotor aspects. Creative thinking behavior is a manifestation of creativity, and creativity refers to the ability to generate new, unique ideas. Related to the importance of behavior on creative thinking, it is necessary to analyze the extent of the development of research on creative thinking behavior. Bibliometric analysis using Publish of Perish (PoP) and VOSviewer for the last 10 years (2014-2023) keyword “creative thinking behavior. The analysis took from the google scholar and scopus databases. Bibliometric analysis is a technique that is realized in 2 categories, namely performance analysis and science mapping. The results of the analysis are stored in a small database used in this study. Since this research only uses Google Scholar and Scopus databases, the research methodology of thinking behavior is also limited. The research results can be expanded by using alternative databases and keyword searches. This study may be useful for future researchers who want to investigate creative thinking behavior.

Do Implicit Learning Deficit and Dyslexia Go Together? An fMRI and Behavioral Study

AbstractWhat is the relationship between literacy skills and implicit learning? To address previous mixed findings, we compared school‐aged readers, typical (CON, n = 54) and with dyslexia (DYS, n = 53), in relation to their performance on a serial reaction time task. For the first time, we also included an isolated spelling deficit group (ISD, n = 30) to control for distinctive effects of reading and spelling deficits. A linear reaction times analysis did not reveal between‐group differences in implicit learning. However, further examination revealed that most CON (65%) and ISD (63%) were implicit learners, whereas most DYS were nonlearners (64%). Brain activity showed differences in early learning phases: CON learners and DYS nonlearners activated the left inferior frontal gyrus (IFG) and left insula more than other groups. Our findings imply that implicit learning is more frequently disrupted in children with dyslexia than in typical readers, and that activation of the left IFG and insula contributes to effective learning in the latter group but it does not in the former.

Generating learning through the crowd

Social media and higher education pedagogy have enjoyed a chequered relationship with significant debates about the efficacy of social media as a site of student centred learning, the manager/host of an individual’s learning trajectory and as a tool of facilitating collaborative learning at scale. This paper presents the findings from the evaluation of Constitution UK, an innovative civic engagement and open learning project run by the London School of Economics and Political Science (UK). This was the lead initiative in an institution-wide shift in pedagogical approach, designed to transform the learning experience through supporting students to be co-producers of knowledge. We argue that some of the behaviours inherent in social media learning (centred on fleeting connections, digital identity and discontinuous engagement) can create the conditions for effective learning through experience and practice, both at scale in open, online modes as well in the face-to-face delivery environment. Challenging the dominant pedagogical approaches of other massive online programmes, Constitution UK brought together a civil community of people engaging in the process of digital citizenship that produced a crowdsourced constitution for the United Kingdom. The learning design of the project successfully engineered both learning and problem solving at scale. The key aspects of the project arising from how social media can facilitate critical thinking, engagement, peer and crowd learning have informed pedagogical change within the mainstream provision of the School for initiatives such as Students as Producers, civic engagement over Brexit and games-based learning.

Conceptual Model of Cross-Boundary Learning in Adventure Tourism

This theoretical study aims to provide a learning model for adventure tour guides to become effective learning facilitators for adventure tourists, who have increasingly sought self-transformation through soft-adventure activities. This objective is drowned by the increasing interest in learning from different cultures among adventure tourists and also the attention to self-transformation through pedagogy, adventure tourism, outdoor adventure education (OAE), and pedagogy are becoming increasingly capable of sharing their theories under the concept of cross-boundary learning. This paper therefore organizes this sharing by considering the applicability of OAE theories to soft adventure tourist based on literature review. It then centers on the community of practice created from interaction between guides and tourists and between tourists as a way of learning, which can lead adventure tourists towards self-transformation, and discusses how guides can create opportunities for such cooperative learning community during their service. As a result, taking a cue from the SECI model, which underlies cross-boundary learning, a conceptual model was presented to implement the facilitator role of the guide in cross-boundary learning effectively. In this conceptual model, the guide’s role extends beyond knowledge transfer. Rather, the results of the study imply that guides are required to be co-creators of knowledge to accompany adventure tourists in their cross-boundary learning. The findings also suggest that guides need to become cross-boundary learners who may also learn from their customers.

Quantized non-volatile nanomagnetic domain wall synapse based autoencoder for efficient unsupervised network anomaly detection

Abstract Anomaly detection in real-time using autoencoders implemented on edge devices is exceedingly challenging due to limited hardware, energy, and computational resources. We show that these limitations can be addressed by designing an autoencoder with low-resolution non-volatile memory-based synapses and employing an effective quantized neural network learning algorithm. We further propose nanoscale ferromagnetic racetracks with engineered notches hosting magnetic domain walls (DW) as exemplary non-volatile memory based autoencoder synapses, where limited state (5-state) synaptic weights are manipulated by spin orbit torque (SOT) current pulses to write different magnetoresistance states. The performance of anomaly detection of the proposed autoencoder model is evaluated on the NSL-KDD dataset. Limited resolution and DW device stochasticity aware training of the autoencoder is performed, which yields comparable anomaly detection performance to the autoencoder having floating-point precision weights. While the limited number of quantized states and the inherent stochastic nature of DW synaptic weights in nanoscale devices are typically known to negatively impact the performance, our hardware-aware training algorithm is shown to leverage these imperfect device characteristics to generate an improvement in anomaly detection accuracy (90.98%) compared to accuracy obtained with floating-point synaptic weights that are extremely memory intensive. Furthermore, our DW-based approach demonstrates a remarkable reduction of at least three orders of magnitude in weight updates during training compared to the floating-point approach, implying significant reduction in operation energy for our method. This work could stimulate the development of extremely energy efficient non-volatile multi-state synapse-based processors that can perform real-time training and inference on the edge with unsupervised data.

PENDAMPINGAN KOMUNITAS BELAJAR SEKOLAH PENGGERAK MELALUI LOKAKARYA REFLEKSI UNTUK MENINGKATKAN KUALITAS PROSES PEMBELAJARAN DI SEKOLAH DASAR

Driving schools, especially elementary schools in Wajo Regency, have several learning communities with different teacher competencies. Most teachers have not fully mastered the differentiation learning process in the Merdeka Belajar curriculum. The aim of this mentoring activity is to assist learning community teachers in planning, organizing and implementing an effective and creative Merdeka curriculum learning process. The method used in this service is a workshop form which includes stages starting from self, collaboration space, contextual demonstration and real action. Meanwhile, the material in this service activity uses teaching materials that have been determined by the Directorate General of Teachers and Education Personnel, Ministry of Education and Culture, Research and Technology. The results of the service show that the learning community that was accompanied has been able to create a plan for real action for a quality learning process that will be implemented in their respective educational units. Apart from that, the participants increasingly understand the function and role of the Learning Community in improving student performance and solving every problem in the Education Unit

Streamlining Ocean Dynamics Modeling with Fourier Neural Operators: A Multiobjective Hyperparameter and Architecture Optimization Approach

Training an effective deep learning model to learn ocean processes involves careful choices of various hyperparameters. We leverage DeepHyper’s advanced search algorithms for multiobjective optimization, streamlining the development of neural networks tailored for ocean modeling. The focus is on optimizing Fourier neural operators (FNOs), a data-driven model capable of simulating complex ocean behaviors. Selecting the correct model and tuning the hyperparameters are challenging tasks, requiring much effort to ensure model accuracy. DeepHyper allows efficient exploration of hyperparameters associated with data preprocessing, FNO architecture-related hyperparameters, and various model training strategies. We aim to obtain an optimal set of hyperparameters leading to the most performant model. Moreover, on top of the commonly used mean squared error for model training, we propose adopting the negative anomaly correlation coefficient as the additional loss term to improve model performance and investigate the potential trade-off between the two terms. The numerical experiments show that the optimal set of hyperparameters enhanced model performance in single timestepping forecasting and greatly exceeded the baseline configuration in the autoregressive rollout for long-horizon forecasting up to 30 days. Utilizing DeepHyper, we demonstrate an approach to enhance the use of FNO in ocean dynamics forecasting, offering a scalable solution with improved precision.

Penerapan Pendekatan Teaching at the Right Level (TaRL) untuk Meningkatkan Hasil Belajar Sejarah Peserta Didik

This research aims to find out how the Teaching at the Right Level (TaRL) approach is implemented as a strategy to improve learning outcomes. TaRL adapts learning to students' level of understanding and encourages problem-based learning. Classroom Action Research (CAR) was conducted at SMAN 14 Medan by applying TaRL. The research method used the Kemmis & McTaggart cycle model consisting of planning, action, observation, and reflection. The research population was students of SMAN 14 Medan, with a sample of grade X-4 students. In cycle I, although there was an increase, the learning outcomes were still not complete. However, in cycle II, there was a significant improvement with 83.33% of students reaching the minimum completion criteria. These results show that the TaRL approach can improve students' understanding of historical material. In addition, the utilization of technology also helps in increasing student participation and expanding the learning resources that students can access. These findings can help education practitioners develop more effective learning strategies that suit students' needs in understanding historical materials.

Community-based learning in medical education: A starting guide and lessons learned

Community-based medical education (CBME) addresses real-world health problems and is characterized by its emphasis on reciprocity and collaboration with community stakeholders. Limited evidence shows that CBME is an effective learning strategy to care for underserved communities. However, medical schools and nursing schools struggle to implement CBME in their curriculum. In this article, we present four practical examples of CBME from medical and nursing schools in Belgium and the Netherlands. By taking the lessons learned derived from these practical examples into account, all students can have an authentic learning experience within the community, which empowers community members and increases their health.

Transitioning to blended learning during COVID‐19: Exploring instructors and adult learners' experiences in three Ghanaian universities

AbstractTeaching and learning in higher education have increasingly become digitalized and associated with innovative pedagogical methodologies over the past decades. Following the sudden onset of the pandemic in March 2020, several studies tended to focus on traditional students' experiences with emergency remote education while literature is scarce on non‐traditional students (or adult learners) pedagogical experiences in Distance Education contexts. Using a qualitative case study approach, this study explored how digital technologies mediated instructors' and adult learners' educational experiences during COVID‐19 and their first‐time experiences with blended learning. Between October 2021 and July 2022, semi‐structured interviews were conducted with 40 adult learners and 20 instructors in three selected public higher education institutions (HEIs) in Ghana. Data analysis followed Braun and Clarke's (2012) thematic analysis approach. The results showed that the transition to blended learning was supported by the adoption and integration of varied virtual online technologies. The themes highlight the positive and negative impacts of technology in mediating the educational experiences of instructors and adult learners in blended learning environments. The themes which reflected both instructors' and adult learners' experiences were enhanced course delivery and pedagogy, competency development, technological issues and poor‐quality pedagogy. Blended learning facilitated by technology could be the ‘new normal learning’ post‐pandemic for adult learners pursuing Distance Education in Ghana. The study recommends the implementation of agile strategies and policies by HEIs to ensure sustainable quality education in distance learning. Practitioner notesWhat is already known about this topic The COVID‐19 pandemic necessitated the adoption of innovative pedagogical approaches in higher education contexts. Technological transformations in information and communication technology (ICT) have enhanced remote teaching in higher education institutions globally. The educational experiences of educators and learners differ in blended learning contexts. What this paper adds Instructors and adult learners' positive experiences with navigating a variety of web‐based technologies during the educational process are negatively impacted by technological difficulties during online instruction. Participants' experiences of blended learning are mixed with a preference for face‐to‐face dimension of blended learning instead of the online dimension. The paper identifies four themes that characterize instructors and adult learners' experiences with technology‐enhanced learning including enhanced course delivery and pedagogy, competency development, technological issues and poor‐quality pedagogy. Implications for practice and/or policy The study provides evidence‐based information on the relevance of digitizing distance education for sustainable development and promotion of lifelong learning opportunities for distance education students. The research recommends that higher education institutions (HEIs) implement agile policies to facilitate a seamless shift to distance learning. HEIs may adopt open distance learning frameworks to streamline 21st century pedagogical and learning practices in distance‐blended learning environments for quality course instruction. The study highlights the potential distance learning modalities that HEIs can consider for Distance Education students to sustain effective quality teaching and learning.

The Influence of Learning Media on the Quality of Progress Office Administration Program

This research aims to describe the use of learning media and the quality of learning in the Office Administration Skills program at Watampone State Vocational School 1 as well as the influence of learning media on the quality of the teaching and learning process there. This research uses experimental research with a population of 208 students there, with a sample of 68 students with an error rate of 10 percent. Data collection was carried out using observation and questionnaires. The data analysis techniques used are descriptive statistics and inferential statistics. The results of this research show that the use of learning media there is classified as good from the aspects of print media, electronic media and reality media. The quality of the teaching and learning process there is assessed as good in terms of active, innovative, creative, effective and enjoyable learning. There is a significant positive influence of the use of learning media on the quality of learning in the office administration skills program at SMK Negeri 1 Watampone, amounting to 17.9 percent

Exploring teachers' understanding and implementation of STEAM: one size does not fit all

IntroductionIn recent years, STEAM (science, technology, engineering, arts, and math) education has become an increasingly popular tool to promote student learning and engagement across disciplines. However, researchers and practitioners continue to note the lack of clarity around definitions of STEAM, its intended purpose, and the nature of discipline integration. In this paper, we explore teachers' understandings and implementation of STEAM as they participated in a teacher professional learning program.MethodsUsing a mixed-methods approach, we analyzed data from 4 years of program implementation with K-12 teachers across multiple schools and districts (n = 61).ResultsParticipants' understanding of STEAM was deepened through their involvement in the professional learning program, with many reflecting on their initial, often shallow, (mis)conceptualizations of STEAM. Using an integration continuum developed by the authors, we identified ways in which integration varied across the STEAM lessons teachers developed, contextualizing these findings within the changing educational landscape of the COVID-19 pandemic, as teachers transitioned between remote, hybrid, and in-person instruction. Participants shared their reflections on the personal, professional, and contextual factors that supported and hindered successful STEAM integration.DiscussionWe found that, when provided with sustained, collaborative pedagogical and material support, teachers could successfully improve their understanding of STEAM and implement STEAM lessons tailored to their grade-level and disciplinary context. Our findings reinforce that teachers need support, autonomy, and flexibility to adopt an approach to STEAM integration that best fits their classroom and school context. We discuss the implications of these findings for researchers and practitioners working to provide effective STEAM teacher professional learning.

Efficientnetv2-RegNet: an effective deep learning framework for secure SDN based IOT network

Efficientnetv2-RegNet: an effective deep learning framework for secure SDN based IOT network

On Chinese learning strategies of learners from Central Asian countries: an analysis of gender, age, and learning duration effects

This research focuses on the Chinese learning strategies employed by learners from Central Asian countries, specifically examining the effects of gender, age, and learning duration. The study aims to identify and analyze the demographic factors that influence the learning approaches of these learners, providing insights for more effective teaching and learning of Chinese as a foreign language. Data are collected through questionnaires and interviews, and statistical analysis is conducted to explore the correlations between gender, age, learning duration, and learning strategies. The learning strategy model adopted in this study provides a comprehensive classification of language learning strategies. The results underscore the significance of incorporating these factors into language education programs, providing valuable insights into the unique needs and challenges encountered by learners from Central Asian countries. The findings indicate that students from Central Asian countries predominantly utilize social strategies, metacognitive strategies, and affective strategies in their Chinese language learning. These are followed by compensatory strategies, cognitive strategies, and memory strategies. While gender does not exert a significant impact on the utilization of Chinese learning strategies, there are discernible variations in memory and affective strategies between males and females. Age does not significantly affect overall learning strategies, but there are notable disparities in compensatory strategies among different age groups. Learning duration has a significant effect on compensation and metacognitive strategies. The correlation between learning duration and overall strategies is significant indicating that learners with different learning durations exhibit notable differences in compensation and metacognitive strategies.

Enhancing Reading Comprehension Proficiency of Grade 10 Students: Input for Supplementary Reading Materials

This present study investigated the reading comprehension levels of the Grade 10 students, their levels of macro and micro reading comprehension skills, and the different reading activities provided to them by the teachers. Considering the effect of the COVID-19 pandemic, the Grade 10 students’ acquisition of reading comprehension skills has been affected. Consequently, assessment of students’ reading comprehension is essential to collect information and make inferences about the level of their reading comprehension skills which would help the Grade 10 teachers determine the underlying problems in students’ reading skills and eventually guide them in the crafting of the most appropriate reading interventions and well-designed remediation plans. The instrument used to assess the student’s level of reading comprehension skills was categorized into two based on Douglas Henry Brown’s micro and macro reading skills and an interview for teacher-respondents on the activities provided to the students in teaching reading. The results of this study indicated that the student’s reading comprehension proficiency was at a proficient level and yet more intensive reading comprehension activities should be given, well-designed supplementary reading materials be provided, and varied and effective learning strategies be employed by the teachers to improve their reading comprehension.

Bridging the Digital Divide: The Role of Online Education in Promoting Equity

This qualitative case study examines the impact of online education in Smith County (a pseudonym), a rural area in the southwestern United States characterized by economic challenges and a large Hispanic population. This study aimed to identify the barriers to and facilitators of educational equity in digital environments. A purposive sampling was utilized to select participants who could provide rich, in-depth insights into the impact of online education within Smith County. In-depth interviews were conducted and focus groups held with students, teachers, and administrators within Smith County. Data analysis was conducted through an iterative process of coding and thematic analysis. The study was informed by the theoretical framework of digital divide and culturally responsive pedagogy. This study delved into the nuanced challenges faced by this community, including technological access, digital literacy, and socio-economic factors. The analysis also explored the pedagogical approaches and support systems that contribute to effective online learning experiences within this rural area. The findings revealed a complex interplay of factors affecting online education engagement and success, highlighting the critical role of tailored support services, culturally responsive teaching, and community partnerships to enhance educational outcomes. The study underscores the importance of understanding local contexts and learner needs in designing and implementing online education programs that specifically address equity. Major conclusions point toward the need for comprehensive strategies that go beyond providing internet access, emphasizing the creation of inclusive, supportive online learning communities. This research contributes to the discourse on educational equity in the digital age, offering insights for educators, policymakers, and online education providers seeking to improve access and outcomes for all learners.

Higher-order embedded learning for heterogeneous information networks and adaptive POI recommendation

Traditional POI recommendation results tend to be homogeneous due to data sparsity, and cannot adapt to the dynamic switching of user locations. Regarding the issues, we propose an adaptive Point of Interest (POI) recommendation algorithm that integrates the user social attributes and POI natural attributes. On the one hand, to fully utilize the complex node interactions and semantic information of heterogeneous networks, an effective higher-order embedding learning framework based on multi-head attention mechanism and multi-task shared features, called UI-HEHo, is presented. It divides hierarchical embedding learning into inter-order and intra-order aggregation to learn richer and more comprehensive representations. Based on UI-HEHo framework, a POI-user social attribute model is constructed to obtain the relational distance between interest points. On the other hand, a POI-natural attributes model based on user preference activity range prediction is constructed to further draw the influence between interest points. Furthermore, a tensor-based dynamic update scheme is designed to reduce the high cost of model updates. Finally, the Radviz graph is utilized to obtain the final recommendation results. Experimental results demonstrate our proposed approach significantly improves the POI recommendation performance, especially achieves 20.7% and 23.8% improvement on the Foursquare and Yelp datasets in terms of the newly defined diversity and novelty metrics.

CSDNet: A Novel Deep Learning Framework for Improved Cataract State Detection

Cataracts, known for lens clouding and being a common cause of visual impairment, persist as a primary contributor to vision loss and blindness, presenting notable diagnostic and prognostic challenges. This work presents a novel framework called the Cataract States Detection Network (CSDNet), which utilizes deep learning methods to improve the detection of cataract states. The aim is to create a framework that is more lightweight and adaptable for use in environments or devices with limited memory or storage capacity. This involves reducing the number of trainable parameters while still allowing for effective learning of representations from data. Additionally, the framework is designed to be suitable for real-time or near-real-time applications where rapid inference is essential. This study utilizes cataract and normal images from the Ocular Disease Intelligent Recognition (ODIR) database. The suggested model employs smaller kernels, fewer training parameters, and layers to efficiently decrease the number of trainable parameters, thereby lowering computational costs and average running time compared to other pre-trained models such as VGG19, ResNet50, DenseNet201, MIRNet, Inception V3, Xception, and Efficient net B0. The experimental results illustrate that the proposed approach achieves a binary classification accuracy of 97.24% (normal or cataract) and an average cataract state detection accuracy of 98.17% (normal, grade 1—minimal cloudiness, grade 2—immature cataract, grade 3—mature cataract, and grade 4—hyper mature cataract), competing with state-of-the-art cataract detection methods. The resulting model is lightweight at 17 MB and has fewer trainable parameters (175, 617), making it suitable for deployment in environments or devices with constrained memory or storage capacity. With a runtime of 212 ms, it is well-suited for real-time or near-real-time applications requiring rapid inference.