• Title/Summary/Keyword: meta-learning

Search Result 316, Processing Time 0.029 seconds

Analysis of Evolutionary Optimization Methods for CNN Structures (CNN 구조의 진화 최적화 방식 분석)

  • Seo, Kisung
    • The Transactions of The Korean Institute of Electrical Engineers
    • /
    • v.67 no.6
    • /
    • pp.767-772
    • /
    • 2018
  • Recently, some meta-heuristic algorithms, such as GA(Genetic Algorithm) and GP(Genetic Programming), have been used to optimize CNN(Convolutional Neural Network). The CNN, which is one of the deep learning models, has seen much success in a variety of computer vision tasks. However, designing CNN architectures still requires expert knowledge and a lot of trial and error. In this paper, the recent attempts to automatically construct CNN architectures are investigated and analyzed. First, two GA based methods are summarized. One is the optimization of CNN structures with the number and size of filters, connection between consecutive layers, and activation functions of each layer. The other is an new encoding method to represent complex convolutional layers in a fixed-length binary string, Second, CGP(Cartesian Genetic Programming) based method is surveyed for CNN structure optimization with highly functional modules, such as convolutional blocks and tensor concatenation, as the node functions in CGP. The comparison for three approaches is analysed and the outlook for the potential next steps is suggested.

A Study on the Aspect of Francophone Korean learners' Use of Listening Strategies (프랑스어권 학습자의 한국어 듣기 전략 사용 양상 연구)

  • Yoon, Saerom;Jang, Younjung
    • Journal of Korean language education
    • /
    • v.29 no.3
    • /
    • pp.145-163
    • /
    • 2018
  • The purpose of this study was to identify the necessity of research for increasing French language learners and to examine their use of listening strategies according to their proficiency as a basic study for their continuous learning and communication skills. In the case of French language Korean learners, both the beginner and intermediate learners used the upper cognitive strategy most frequently. However, the cognitive strategy, which has been mentioned as a frequently used strategy in previous studies, was found to be the least used in this study. This finding can be attributed to differences in mores and mastery of prior studies and research subjects. The cognitive strategy was lower in both the beginner and intermediate levels, but the level of use increased significantly in the intermediate level compared to the beginner level, showing only statistically significant differences in the usage patterns according to the proficiency level among the four listening strategies.

A multi-objective decision making model based on TLBO for the time - cost trade-off problems

  • Eirgash, Mohammad A.;Togan, Vedat;Dede, Tayfun
    • Structural Engineering and Mechanics
    • /
    • v.71 no.2
    • /
    • pp.139-151
    • /
    • 2019
  • In a project schedule, it is possible to reduce the time required to complete a project by allocating extra resources for critical activities. However, accelerating a project causes additional expense. This issue is addressed by finding optimal set of time-cost alternatives and is known as the time-cost trade-off problem in the literature. The aim of this study is to identify the optimal set of time-cost alternatives using a multiobjective teaching-learning-based optimization (TLBO) algorithm integrated with the non-dominated sorting concept and is applied to successfully optimize the projects ranging from a small to medium large projects. Numerical simulations indicate that the utilized model searches and identifies optimal / near optimal trade-offs between project time and cost in construction engineering and management. Therefore, it is concluded that the developed TLBO-based multiobjective approach offers satisfactorily solutions for time-cost trade-off optimization problems.

Metaverse Platform Design for Strengthening Gender Sensitivity of MZ Generation

  • Kim, Sea Woo;Na, Eun Gyung
    • International journal of advanced smart convergence
    • /
    • v.11 no.3
    • /
    • pp.79-84
    • /
    • 2022
  • Due to a series of online sex crimes cases and online class conversions caused by the spread of the coronavirus, alternatives to sex education in schools are urgently required. As a result of this study, the metaverse sex education platform was designed. Using this platform, learners are expected to cultivate correct adult awareness and digital citizenship. Within the metaverse platform, learners can participate more actively in learning. Instead of exposing one's name and face in a place dealing with sensitive gender issues, one can participate in education through his or her decorated avatar and participate in education much more actively than face-to-face education and express one's opinion through chat. In addition, education by level can be received regardless of time and place, which can have the effect of bridging the educational gap between urban and rural areas. In this paper, we propose a new sex education platform without time and space constraints by utilizing metaverse.

Adaptive Neuro-fuzzy-based modeling of exhaust emissions from dual-fuel engine using biodiesel and producer gas

  • Prabhakar Sharma;Avdhesh Kr Sharma
    • Advances in Energy Research
    • /
    • v.8 no.3
    • /
    • pp.175-184
    • /
    • 2022
  • The dual-fuel technology, which uses gaseous fuel as the main fuel and liquid as the pilot fuel, is an appealing technology for reducing the exhaust emissions. The current study proposes emission models based on ANFIS for a dual-fuel using producer gas (PG)-diesel engine. Emissions measurements were taken at different engine load levels and fuel injection timings. The proposed model predictions were examined using statistical methods. With R2 values in the range of 0.9903 to 0.9951, the established ANFIS model was found to be consistently robust in predicting emission characteristics. The mean absolute percentage deviate in range 1.9 to 4.6%, and mean squared error varies in range 0.0018 to 13.9%. The evaluation of the ANFIS model developed shows a reliable claim of intrinsic sensitivity, strength, and outstanding generalization. The presented meta-model can be used to simulate the engine's operation in order to create an efficient control tool.

A Study and Design of Meta-Learning for Intelligent Tutoring System (지능형 튜터링 시스템을 위한 메타러닝 설계 연구)

  • Hong, Seong-Yong
    • Proceedings of the Korea Information Processing Society Conference
    • /
    • 2010.11a
    • /
    • pp.429-431
    • /
    • 2010
  • 인터넷과 정보기술의 발전으로 최근 이러닝 시스템을 포함한 다양한 학습 시스템이 연구 발전되고 있다. 학습자의 관점에서는 학습의 형태 혹은 학습자의 학습 패턴등을 분석하여 지능적인 학습시스템으로 발전하고 있으며, 교수자의 관점에서는 교수학습 모델 연구와 학습 컨텐츠 계발 방법론 연구 등이 활발하게 이루어지고 있다. 본 논문에서는 지능형 튜터링 시스템을 위한 메타러닝 설계 연구를 제안하였다. 메타러닝은 학습자 자신이 어떤 특성을 가지고 어떻게 학습하는지에 대해 학습할 수 있는 방법을 설명한다. 동일한 학습내용을 같은 순서 혹은 같은 방법으로 학습하는 것은 서로 다른 학습자에게 동일한 학습 결과를 나타낼 수 없기 때문에 개인 맞춤형 학습 서비스 형태를 필요로 한다. 따라서 본 연구에서는 메타러닝 설계를 기반으로 지능형 튜터링 시스템을 개발 할 수 있는 방법을 설명하고자 한다. 향후 본 논문에 설계를 기반으로 지능형 튜터링 플랫폼을 표준으로 개발하여 국제적 표준의 ITS(Intelligent Tutoring System)로 발전되기를 기대한다.

Classification for Imbalanced Breast Cancer Dataset Using Resampling Methods

  • Hana Babiker, Nassar
    • International Journal of Computer Science & Network Security
    • /
    • v.23 no.1
    • /
    • pp.89-95
    • /
    • 2023
  • Analyzing breast cancer patient files is becoming an exciting area of medical information analysis, especially with the increasing number of patient files. In this paper, breast cancer data is collected from Khartoum state hospital, and the dataset is classified into recurrence and no recurrence. The data is imbalanced, meaning that one of the two classes have more sample than the other. Many pre-processing techniques are applied to classify this imbalanced data, resampling, attribute selection, and handling missing values, and then different classifiers models are built. In the first experiment, five classifiers (ANN, REP TREE, SVM, and J48) are used, and in the second experiment, meta-learning algorithms (Bagging, Boosting, and Random subspace). Finally, the ensemble model is used. The best result was obtained from the ensemble model (Boosting with J48) with the highest accuracy 95.2797% among all the algorithms, followed by Bagging with J48(90.559%) and random subspace with J48(84.2657%). The breast cancer imbalanced dataset was classified into recurrence, and no recurrence with different classified algorithms and the best result was obtained from the ensemble model.

Arrhythmia classification based on meta-transfer learning using 2D-CNN model (2D-CNN 모델을 이용한 메타-전이학습 기반 부정맥 분류)

  • Kim, Ahyun;Yeom, Sunhwoong;Kim, Kyungbaek
    • Proceedings of the Korea Information Processing Society Conference
    • /
    • 2022.11a
    • /
    • pp.550-552
    • /
    • 2022
  • 최근 사물인터넷(IoT) 기기가 활성화됨에 따라 웨어러블 장치 환경에서 장기간 모니터링 및 수집이 가능해짐에 따라 생체 신호 처리 및 ECG 분석 연구가 활성화되고 있다. 그러나, ECG 데이터는 부정맥 비트의 불규칙적인 발생으로 인한 클래스 불균형 문제와 근육의 떨림 및 신호의 미약등과 같은 잡음으로 인해 낮은 신호 품질이 발생할 수 있으며 훈련용 공개데이터 세트가 작다는 특징을 갖는다. 이 논문에서는 ECG 1D 신호를 2D 스펙트로그램 이미지로 변환하여 잡음의 영향을 최소화하고 전이학습과 메타학습의 장점을 결합하여 클래스 불균형 문제와 소수의 데이터에서도 빠른 학습이 가능하다는 특징을 갖는다. 따라서, 이 논문에서는 ECG 스펙트럼 이미지를 사용하여 2D-CNN 메타-전이 학습 기반 부정맥 분류 기법을 제안한다.

A Survey on Methodology of Meta-Learning (메타 러닝과 방법론 연구 동향)

  • Hoon Ji;Yeon-Joon Lee
    • Proceedings of the Korea Information Processing Society Conference
    • /
    • 2023.05a
    • /
    • pp.665-666
    • /
    • 2023
  • 딥러닝은 인간이 탐지하기 어려운 데이터의 특징 및 패턴을 인지하고, 이들을 학습하여 데이터를 분류 및 예측할 수 있는 기술이다. 그러나 딥러닝 모델을 잘 학습시키기 위해서는 고품질의 대용량 데이터와 이들을 처리할 수 있는 방대한 컴퓨터 자원이 요구되는 것이 일반적이다. 따라서 소량의 데이터만이 존재하는 분야나 컴퓨터 자원이 한정되어 있는 상황에서는 딥러닝을 적용하기 어렵다. 본 논문에서는, 소량의 데이터로도 모델을 자신들의 태스크에 맞게 최적화시킬 수 있는 메타러닝에 대해 소개하고, 메타 러닝 기법들의 방향에 따른 Metric-Based, Model-Based 및 Optimization 기반 모델들에 대해 소개하고, 앞으로 나아가야 할 연구 방향에 대해 제시한다.

The Effect of Writing a Weekly Report on the Self-directed Learning, Attitude toward science, and Academic achievement (주 단위 보고서 작성이 자기 주도적 학습 능력과 과학에 대한 태도 및 학업 성취도에 미치는 영향)

  • Kim, Mijung;Woo, AeJa
    • Journal of Science Education
    • /
    • v.39 no.2
    • /
    • pp.165-179
    • /
    • 2015
  • In this study, the effects of writing a weekly report on the students' self-directed learning, the attitudes toward science, and the academic achievements were examined. Two hundred and three students, second graders of a high school participated. Experimental group performed writing a weekly report, while the comparative group performed regular science lessons. The results of this study are as follows: First, MSLQ test showed that there was statistically significant difference in the self-directed learning skills(p<.05). For sub-factors of motivation region, such as internal goals, extrinsic goals, learning beliefs, task value, and self-efficacy and for sub-factors of learning strategy region, such as meta-cognition, peer learning, time management, critical thinking, and demonstrations showed statistically significant results. Second, TOSRA test showed that there was no statistically significant difference in the attitudes toward science (p>.05). However, for sub-factors, such as scientific inquiry and joy to science class showed statistically significant results. Third, there was no statistically significant difference in the academic achievement in Chemistry I class (p>.05). However, top and low achievement level showed statistically significant results.

  • PDF