• International Journal of Internet, Broadcasting and Communication
• /
• 제16권3호
• /
• pp.262-273
• /
• 2024

We provide a detailed analysis of the data processing and model training process for vulnerability classification using Transformer-based language models, especially sentence text-to-text transformers (ST5)-XXL and XLNet. The main purpose of this study is to compare the performance of the two models, identify the strengths and weaknesses of each, and determine the optimal learning rate to increase the efficiency and stability of model training. We performed data preprocessing, constructed and trained models, and evaluated performance based on data sets with various characteristics. We confirmed that the XLNet model showed excellent performance at learning rates of 1e-05 and 1e-04 and had a significantly lower loss value than the ST5-XXL model. This indicates that XLNet is more efficient for learning. Additionally, we confirmed in our study that learning rate has a significant impact on model performance. The results of the study highlight the usefulness of ST5-XXL and XLNet models in the task of classifying security vulnerabilities and highlight the importance of setting an appropriate learning rate. Future research should include more comprehensive analyzes using diverse data sets and additional models.

• KIEAE Journal
• /
• 제17권5호
• /
• pp.95-100
• /
• 2017

Purpose: The purpose of this study is to investigate the prior art based on deep learning to objectively calculate the metabolic rate which is the subjective factor for the PMV optimum control and to make a plan for future research based on this study. Methods: For this purpose, the theoretical and technical review and applicability analysis were conducted through various documents and data both in domestic and foreign. Results: As a result of the prior art research, the machine learning model of artificial neural network and deep learning has been used in various fields such as speech recognition, scene recognition, and image restoration. As a representative case, OpenCV Background Subtraction is a technique to separate backgrounds from objects or people. PASCAL VOC and ILSVRC are surveyed as representative technologies that can recognize people, objects, and backgrounds. Based on the results of previous researches on deep learning based on metabolic rate for occupational metabolic rate, it was found out that basic technology applicable to occupational metabolic rate calculation technology to be developed in future researches. It is considered that the study on the development of the activity quantity calculation model with high accuracy will be done.

• Educational Technology International
• /
• 제11권1호
• /
• pp.27-46
• /
• 2010

The purpose of this paper is to apply the newly developed SICAT teaching and learning model to the actual scene of teaching and learning and draw a point of discussion for utilizing teaching and learning model, by uncovering the satisfaction of students and the inhibiting/facilitating elements when using the model. SICAT(Scientific Inquiry and Creative Activity with Technology; from here on SICAT), a teaching and learning model custom-built for engineering education, was developed, as more and more people paid attention to the demand for creative engineers. It was developed from the basis of PBL(Problem Based Learning), includes three sub-types which can be applied to the actual theory, design, and experimentation fields within engineering education. The three sub-types, which are ARDA(Analysis-Reasoning Activity & Discussion-Argumentation Activity), CoCD (Collaboration Activity & Capstone Design Activity), and ReSh(Reflection Activity & Sharing Activity), respectively support deductive and argumentation activities, creative design and collaboration activities, and retrospection and sharing activities. However, no research has been conducted to investigate whether or not there are inhibiting or facilitating elements in the application procedure, or what the rate of satisfaction for students is, when applying the SICAT model, which was newly developed to innovate existing engineering education, to the actual site of teaching and learning. Therefore, this research applied three types of SICAT teaching and learning models to the theory, design, and experimentation classes at the department of materials science and engineering at Hanyang University for eight weeks. After application, the students, teachers and tutors were surveyed and interviewed, and then the results analyzed in order to uncover inhibiting/facilitating elements and the rate of satisfaction. The satisfaction rate of students from the SICAT teaching and learning model was 3.78(in a perfect score of 5: The A type-3.65, The C type-3.80, The R type-3.90), and inhibiting/facilitating elements were drawn from the aspects of learning activities, support system. In conclusion, they can be contributed for implications of SICAT teaching and learning model universal use at engineering education in University.

• 한국어교육
• /
• 제29권4호
• /
• pp.93-112
• /
• 2018

The purpose of this study is to deal with vocabulary in KFL. As a result of this study, learners learned vocabulary on average 43 points through contextual inference and introduction of the class to activate background knowledge. In particular, the implicit method showed the highest learning rate of 52 points, and the thematic method had a 41 point-learning rate. In contrast, the semantic method was the lowest with a 25 point-learning rate. There was no significant difference in the improvement rate of upper vocabulary learners, but in the case of the lower learner, there was significant difference in the improvement rate. The difference was not significant in the post-test relative gain rate of upper learners, but there was significant in lower learners. In the delayed test relative gain rate, the difference was significant in all groups. There was correlation between vocabulary difficulty and score, but there was no correlation with the thematic method. And there was no correlation between vocabulary difficulty, improvement rate and relative gain rate in all three classes. However, content understanding, lexical grade, improvement rate, and relative gain rate showed a significant correlation.

• 공학교육연구
• /
• 제21권1호
• /
• pp.37-43
• /
• 2018

The purpose of this study is to investigate the effects of peer tutoring and feedback on academic learning in university mathematics. We compared subject satisfaction and academic achievement between the test group and the control group. We classified the test group(82 participants) and the control group(134 non-participants) and then applied peer tutoring and feedback to the test group. The rest of the environment was the same except for participation in the program. According to results, it was confirmed that the subject satisfaction were significantly higher(significance level .05) in the test group, where the subject satisfaction were learning objectives and expectation, learning satisfaction, and learning effect. Furthermore, in the change of academic achievement, the rate of decrease was lower and the rate of increase was higher in the test group than the control group. The satisfaction of participants was 4.33(Likert scale 5), and this trend tended to be same regardless of gender, high school course, or admission process.

• Journal of Information Processing Systems
• /
• 제16권5호
• /
• pp.1048-1063
• /
• 2020

In this study, we propose a reinforcement learning agent to control the data transmission rates of nodes in carrier sensing multiple access with collision avoidance (CSMA/CA)-based wireless networks. We design a reinforcement learning (RL) agent, based on Q-learning. The agent learns the environment using the timeout events of packets, which are locally available in data sending nodes. The agent selects actions to control the data transmission rates of nodes that adjust the modulation and coding scheme (MCS) levels of the data packets to utilize the available bandwidth in dynamically changing channel conditions effectively. We use the ns3-gym framework to simulate RL and investigate the effects of the parameters of Q-learning on the performance of the RL agent. The simulation results indicate that the proposed RL agent adequately adjusts the MCS levels according to the changes in the network, and achieves a high throughput comparable to those of the existing data transmission rate adaptation schemes such as Minstrel.

• 한국정보통신학회논문지
• /
• 제7권5호
• /
• pp.853-862
• /
• 2003

본 논문에서는 퍼지 이론을 이용한 학습 평가 방법을 제안하였다. 제안된 학습 평가 방법은 정보처리 데이터베이스 과목에 대한 기출문제의 출제 빈도수를 3등급으로 분류하고 이것을 중요도라 정의하였다. 학습 중요도에 따른 학습 횟수에 대한 퍼지 소속도와 형성평가 점수에 대한 퍼지 소속도를 각각 9개의 퍼지 추론 규칙에 적용하여 학습 이해도를 평가하였다. 최종적인 학습 평가는 각 장별 학습 이해도에 대한 퍼지 등급과 총괄평가 점수에 대한 소속도를 이용하여 퍼지 추론규칙에 적용하고 비퍼지화하여 평가하였다. 제시된 퍼지 이론을 이용한 학습 평가 방법은 학습자가 스스로 학습한 내용을 진단 할 수 있도록 도와주며, 학습목표의 성취여부를 종합적이고 객관적으로 판단할 수 있는 방법을 제공한다.

• 정보학연구
• /
• 제8권1호
• /
• pp.119-124
• /
• 2005

Learning procedure in the neural network is updating of weights between neurons. Unadequate initial learning coefficient causes excessive iterations of learning process or incorrect learning results and degrades learning efficiency. In this paper, adaptive learning algorithm is proposed to increase the efficient in the learning algorithms of Kohonens Self-Organization Neural networks. The algorithm updates the weights adaptively when learning procedure runs. To prove the efficiency the algorithm is experimented to clustering of the random weight. The result shows improved learning rate about 42~55% ; less iteration counts with correct answer.

• 한국정보통신학회논문지
• /
• 제8권2호
• /
• pp.464-470
• /
• 2004

본 논문에서는 역전파 알고리즘의 성능 개선을 위해 퍼지 제어 시스템을 적용하여 학습률을 자동으로 조정하는 개선된 역전파 알고리즘을 제안한다. 제안된 방법은 목표값과 출력값의 차이에 대한 절대값이 $\varepsilon$ 보다 적거나 같으면 정확성으로 분류하고 크면 부정확성으로 분류한다. 정확성과 부정확성의 개수를 퍼지 제어 시스템에 적용하여 학습률을 동적으로 조정한다. 제안된 방법을 XOR 문제와 숫자 패턴 분류에 적용하여 실험한 결과, 기존의 역전파 알고리즘, 모멘텀 방식, Jacob의 delta-bar-delta 방식보다 성능이 개선됨을 확인하였다.

• 한국멀티미디어학회논문지
• /
• 제25권6호
• /
• pp.878-885
• /
• 2022

Research in the field of computer vision based on deep learning is being actively conducted. However, deep learning-based models have vulnerabilities in adversarial attacks that increase the model's misclassification rate by applying adversarial perturbation. In particular, in the case of FGSM, it is recognized as one of the effective attack methods because it is simple, fast and has a considerable attack success rate. Meanwhile, as one of the efforts to visualize deep learning models, Grad-CAM enables visual explanation of convolutional neural networks. In this paper, I propose a method to generate adversarial examples with high attack success rate by applying Grad-CAM to FGSM. The method chooses fixels, which are closely related to labels, by using Grad-CAM and add perturbations to the fixels intensively. The proposed method has a higher success rate than the FGSM model in the same perturbation for both targeted and untargeted examples. In addition, unlike FGSM, it has the advantage that the distribution of noise is not uniform, and when the success rate is increased by repeatedly applying noise, the attack is successful with fewer iterations.