scholarly journals Stream-Based Extreme Learning Machine Approach for Big Data Problems

2015 ◽  
Vol 2015 ◽  
pp. 1-17 ◽  
Author(s):  
Euler Guimarães Horta ◽  
Cristiano Leite de Castro ◽  
Antônio Pádua Braga
Keyword(s):  
Big Data ◽  
Active Learning ◽  
Learning Strategies ◽  
Data Streams ◽  
High Performance ◽  
Hebbian Learning ◽  
Active Learning Method ◽  
Learning Machine ◽  
Hidden Layer ◽  
Few Data

Big Data problems demand data models with abilities to handle time-varying, massive, and high dimensional data. In this context, Active Learning emerges as an attractive technique for the development of high performance models using few data. The importance of Active Learning for Big Data becomes more evident when labeling cost is high and data is presented to the learner via data streams. This paper presents a novel Active Learning method based on Extreme Learning Machines (ELMs) and Hebbian Learning. Linearization of input data by a large size ELM hidden layer turns our method little sensitive to parameter setting. Overfitting is inherently controlled via the Hebbian Learning crosstalk term. We also demonstrate that a simple convergence test can be used as an effective labeling criterion since it points out to the amount of labels necessary for learning. The proposed method has inherent properties that make it highly attractive to handle Big Data: incremental learning via data streams, elimination of redundant patterns, and learning from a reduced informative training set. Experimental results have shown that our method is competitive with some large-margin Active Learning strategies and also with a linear SVM.

scholarly journals A comprehensive active learning method for multiclass imbalanced data streams with concept drift

2021 ◽  
Vol 215 ◽  
pp. 106778
Author(s):  
Weike Liu ◽  
Hang Zhang ◽  
Zhaoyun Ding ◽  
Qingbao Liu ◽  
Cheng Zhu
Keyword(s):  
Active Learning ◽  
Data Streams ◽  
Concept Drift ◽  
Imbalanced Data ◽  
Learning Method ◽  
Active Learning Method

scholarly journals K-Nearest Robust Active Learning on Big Data and Application in Epitope Prediction

2021 ◽  
Vol 2021 ◽  
pp. 1-9
Author(s):  
Tianchi Lu
Keyword(s):  
Big Data ◽  
Data Analysis ◽  
Active Learning ◽  
Learning Algorithm ◽  
Epitope Prediction ◽  
Big Data Analysis ◽  
Epitope Region ◽  
Specific Antibodies ◽  
Active Learning Method

B-cells that induce antigen-specific immune responses in vivo produce large numbers of antigen-specific antibodies by recognizing subregions (epitopes) of antigenic proteins, in which they can inhibit the function of antigen protein. Epitope region prediction facilitates the design and development of vaccines that induce the production of antigen-specific antibodies. There are many diseases which are difficult to treat without vaccines. And the COVID-19 has destroyed many people’s lives. Therefore, making vaccines to COVID-19 is very important. Making vaccines needs a large number of experiments to get labeled targets. However, obtaining tremendous labeled data from experiments is a challenge for humans. Big data analysis has proposed some solutions to deal with this challenge. Big data technology has developed very fast and has been applied in many areas. In the bioinformatics area, big data analysis solves a large number of problems, particularly in the area of active learning. Active learning is a method of building more predictive models with less labeled data. Active learning establishes models with less data by asking the oracle (human) for the most valuable samples to train models. Hence, active learning’s application in making vaccines is meaningful that the scientists do not need to do tremendous experiments. This paper proposed a more robust active learning method based on uncertainty sampling and K-nearest density and applies it to the vaccine manufacture. This paper evaluates the new algorithm with accuracy and robustness. In order to evaluate the robustness of active learners, a new robustness index is designed in this paper. And this paper compares the new algorithm with a pool-based active learning algorithm, density-weighted active learning algorithm, and traditional machine learning algorithm. Finally, the new algorithm is applied to epitope prediction of B-cell data, which is significant to making vaccines.

scholarly journals Double-Criteria Active Learning for Multiclass Brain-Computer Interfaces

2020 ◽  
Vol 2020 ◽  
pp. 1-13
Author(s):  
Qingshan She ◽  
Kang Chen ◽  
Zhizeng Luo ◽  
Thinh Nguyen ◽  
Thomas Potter ◽  
...  
Keyword(s):  
Active Learning ◽  
Training Data ◽  
Computer Interfaces ◽  
Technological Advances ◽  
Training Samples ◽  
Eeg Data ◽  
Classifier Performance ◽  
Active Learning Method ◽  
Learning Machine ◽  
Elm Classifier

Recent technological advances have enabled researchers to collect large amounts of electroencephalography (EEG) signals in labeled and unlabeled datasets. It is expensive and time consuming to collect labeled EEG data for use in brain-computer interface (BCI) systems, however. In this paper, a novel active learning method is proposed to minimize the amount of labeled, subject-specific EEG data required for effective classifier training, by combining measures of uncertainty and representativeness within an extreme learning machine (ELM). Following this approach, an ELM classifier was first used to select a relatively large batch of unlabeled examples, whose uncertainty was measured through the best-versus-second-best (BvSB) strategy. The diversity of each sample was then measured between the limited labeled training data and previously selected unlabeled samples, and similarity is measured among the previously selected samples. Finally, a tradeoff parameter is introduced to control the balance between informative and representative samples, and these samples are then used to construct a powerful ELM classifier. Extensive experiments were conducted using benchmark and multiclass motor imagery EEG datasets to evaluate the efficacy of the proposed method. Experimental results show that the performance of the new algorithm exceeds or matches those of several state-of-the-art active learning algorithms. It is thereby shown that the proposed method improves classifier performance and reduces the need for training samples in BCI applications.

scholarly journals Active Learning Strategies for Diverse Learning Styles: Simulations Are Only One Method

2013 ◽  
Vol 46 (04) ◽  
pp. 818-822 ◽  
Author(s):  
Pam Bromley
Keyword(s):  
Active Learning ◽  
Learning Styles ◽  
Learning Strategies ◽  
Learning Style ◽  
Group Discussion ◽  
Learning Preferences ◽  
Diverse Learning ◽  
Active Learning Method ◽  
Active Learning Strategies ◽  
Science Instructors

AbstractAlthough political science instructors increasingly recognize the advantages of incorporating active learning activities into their teaching, simulations remain the discipline's most commonly used active learning method. While certainly a useful strategy, simulations are not the only way to bring active learning into classrooms. Indeed, because students have diverse learning styles—comprised of their discrete learning preferences—engaging them in a variety of ways is important. This article explores six active learning techniques: simulations, case studies, enhanced lectures, large group discussion, small group work, and in-class writing. Incorporating these activities into an introductory, writing-intensive seminar on globalization and surveying students about their engagement with course activities, I find that different activities appeal to students with different learning preferences and that simulations are not students most preferred activity. Bringing a broader range of active learning strategies into courses can improve teaching for all students, no matter their learning style.

Faculty Adoption of Active Learning Strategies via Paired Teaching: Conclusions From Two Science Departments

2019 ◽  
Vol 049 (01) ◽  
Author(s):  
Linda Strubbe ◽  
Jared Stang ◽  
Tara Holland ◽  
Sarah Bean Sherman ◽  
Warren Code
Keyword(s):  
Active Learning ◽  
Learning Strategies ◽  
Active Learning Strategies ◽  
Science Departments
Sign in / Sign up

Export Citation Format

Share Document