• Journal of Information Processing Systems
• /
• v.17 no.3
• /
• pp.441-452
• /
• 2021

Federated learning provides an efficient integrated model for distributed data, allowing the local training of different data. Meanwhile, the goal of multi-task learning is to simultaneously establish models for multiple related tasks, and to obtain the underlying main structure. However, traditional federated multi-task learning models not only have strict requirements for the data distribution, but also demand large amounts of calculation and have slow convergence, which hindered their promotion in many fields. In our work, we apply the rank constraint on weight vectors of the multi-task learning model to adaptively adjust the task's similarity learning, according to the distribution of federal node data. The proposed model has a general framework for solving optimal solutions, which can be used to deal with various data types. Experiments show that our model has achieved the best results in different dataset. Notably, our model can still obtain stable results in datasets with large distribution differences. In addition, compared with traditional federated multi-task learning models, our algorithm is able to converge on a local optimal solution within limited training iterations.

• East Asian mathematical journal
• /
• v.38 no.4
• /
• pp.497-516
• /
• 2022

In this study, we tried to devise a method to activate meta-affect in the aspect of supporting mathematics teaching and learning according to the need to find specific strategies and teaching and learning methods to activate learners' meta-affect in mathematics subjects, which are highly influenced by psychological factors. To this end, the definitional and conceptual elements of meta-affect which are the basis of this study, were identified from previous studies. Reflecting these factors, a teaching and learning model that activates meta-affect was devised, and a meta-affect activation strategy applied in the model was constructed. The mathematics teaching and learning model that activates meta-affect developed in this study was refined by verifying its suitability and convenience in the field through expert advice and application of actual mathematics classes. The developed model is meaningful in that it proposed a variety of practical teaching and learning methods that activate the meta-affect of learners in a mathematical learning situation.

• International Journal of Internet, Broadcasting and Communication
• /
• v.15 no.2
• /
• pp.133-143
• /
• 2023

In this study, we propose a big data set of strawberry pests collected directly for diagnosis model learning and an automatic pest diagnosis model architecture based on deep learning. First, a big data set related to strawberry pests, which did not exist anywhere before, was directly collected from the web. A total of more than 12,000 image data was directly collected and classified, and this data was used to train a deep learning model. Second, the deep-learning-based automatic pest diagnosis module is a module that classifies what kind of pest or disease corresponds to when a user inputs a desired picture. In particular, we propose a model architecture that can optimally classify pests based on a convolutional neural network among deep learning models. Through this, farmers can easily identify diseases and pests without professional knowledge, and can respond quickly accordingly.

• Journal of Platform Technology
• /
• v.11 no.4
• /
• pp.80-88
• /
• 2023

To construct a learning model in the field of artificial intelligence, a dataset should be collected and be delivered to the central server where the learning model is constructed. Federated learning is a machine learning method building a global learning model without transmitting data located in a client side in a collaborative manner. It can be used to protect privacy, and after constructing a local trained model on individual clients, the parameters of the local model are aggregated centrally to update the global model. In this paper, we reuse the existing learning parameter to improve federated learning, describe incremental federated learning. For this work, we do experiments using the federated learning framework named Flower, and evaluate the experiment results with regard to elapsed time and precision when executing optimization algorithms.

• Journal of the Korean Society of Safety
• /
• v.34 no.6
• /
• pp.50-57
• /
• 2019

In this study, various types of deep learning models that have been proposed recently are classified according to data input / output types and analyzed to find the deep learning model suitable for constructing a crack detection model. First the deep learning models are classified into image classification model, object segmentation model, object detection model, and instance segmentation model. ResNet-101, DeepLab V2, Faster R-CNN, and Mask R-CNN were selected as representative deep learning model of each type. For the comparison, ResNet-101 was implemented for all the types of deep learning model as a backbone network which serves as a main feature extractor. The four types of deep learning models were trained with 500 crack images taken from real concrete structures and collected from the Internet. The four types of deep learning models showed high accuracy above 94% during the training. Comparative evaluation was conducted using 40 images taken from real concrete structures. The performance of each type of deep learning model was measured using precision and recall. In the experimental result, Mask R-CNN, an instance segmentation deep learning model showed the highest precision and recall on crack detection. Qualitative analysis also shows that Mask R-CNN could detect crack shapes most similarly to the real crack shapes.

• Communications of Mathematical Education
• /
• v.29 no.2
• /
• pp.197-214
• /
• 2015

This study is to design Flipped Learning classrooms of learner-centered education on the pre-service math teachers education. The study aims to explore the feasibility of teaching and learning method. To achieve the objectives of the study was to explore the teaching and learning model. Flipped learing classroom design includes a main step of a typical process of teaching system. And we designed the model based on the ADDIE Model. This model contains the design steps and the Flipped learning component of the teaching and learning design model. Designed classroom presented in three steps that are before classroom, during classroom and after classroom.

• Journal of Information Technology Applications and Management
• /
• v.21 no.4
• /
• pp.141-156
• /
• 2014

This study intended to evaluate the viability of cyber education in terms of learning satisfaction and learning achievement. The study integrated two research streams such as social presence model and learning environment model. Where the learning environment model emphasizes the components of learning aids, social presence model considers more deeply the relationships among peers and with instructors. These two research streams have been considered relatively independently. The study integrated these ideas and measured their reliabilities and validities. The results demonstrate that the two constructs are relevantly independent and both of these constructs are very important considerations for the success of cyber education. The study concludes that cyber education 2.0 requires more social presence factors than the learning environment factors such as technological development or new equipments.

• Journal of Fisheries and Marine Sciences Education
• /
• v.25 no.3
• /
• pp.643-660
• /
• 2013

The numbers of learning underachievers in college are gradually increasing. As a result, the need for extracurricular programs to increase learning in college is also growing. The purpose of this study was to analyze factors of learning difficulty and develop a model of learning consulting for college underachievers. This study also aimed to evaluate this model's validity. Using both 56 subscription forms of college underachievers and three sets of focus group interviews at B university, we found that students had difficulties in goal and career setting, management of grades and tests, learning methods, time management, failure overcome ability, lack of learning habit sustaining power and learning motivation, and so on. We developed a model of learning consulting for college underachievers based on these factors and applied the model to evaluate it's validity, testing it on 31 underachievers currently enrolled in college, five times every week. Let we say in conclusion that this model of learning consulting had positive effects on changing college underachiever's character, emotion, motivation, and behavior towards learning.

• Journal of The Korean Association For Science Education
• /
• v.32 no.8
• /
• pp.1333-1344
• /
• 2012

This study aimed to enhance creative problem solving skill by using the Creative Problem Solving (CPS) learning model which was developed based on creative problem solving approach and five essential features of inquiry. The key strategy of the CPS learning model is using real life problem situations to provide students opportunities to practice creative problem solving skill through 5 learning steps: engaging, problem exploring, solutions creating, plan executing, and concepts examining. The science content used for examining the CPS learning model was "matter and properties of matter" that consists of 3 learning units: Matter, Solution, and Acid-Base Solution. The process to assess the effectiveness of the learning model used the experimental design of the Pretest-Posttest Control-Group Design. Seventh grade-students in the experimental group learned by the CPS learning model. At the same time, students at the same grade level in the control group learned by conventional learning model. The learning models and students' prior knowledge levels were served as the independent variables. The creative problem solving skill was classified in to 4 aspects in: fluency, flexibility, originality, and reasoning. The results indicated that in all aspects, the students' mean scores of creative problem solving between students in experimental group and control group were significantly different at the .05 level. Also, the progression of students' creative problem solving skills was found highly progressed at the later instructional periods. When comparing the creative problem solving scores between groups of students with different levels of prior knowledge, the differences of their creative problem solving scores were founded at .05 level. The findings of this study confirmed that the CPS learning model is effective in enhancing the students' creative problem solving skill.

• Journal of Korea Society of Digital Industry and Information Management
• /
• v.11 no.1
• /
• pp.135-144
• /
• 2015

This study aims to propose an evaluation model that enables cooperative learning using e-Learning system. Even if the teacher and the student are not in the same place at the same time, the team project deliverable submitted by the student to the online system can be viewed by the teacher, enabling the teacher to assess the student not only based on the project but also in many other aspects. The proposed e-learning cooperative learning model allows the development of assessment factors, using such factors in assessment of the student's activities which are performed through the e-learning system, and the feedback of the results to the student so that the student is further motivated for learning. The teacher performs a comprehensive assessment of such factors, which is considered in conjunction with the student's assessment. Implementing the cooperative learning model proposed in this study in various e-learning systems such as Moodle is expected to motivate the student for learning, produces better cooperative learning results, provides greater convenience of assessment to the teacher, and improves fairness of assessment by showing the student's activities in real time.