Cited 0 time in webofscience Cited 0 time in scopus

Deep machine learning for Human deep learning

Title
Deep machine learning for Human deep learning
Authors
Lee, Doo Seok
DGIST Authors
Lee, Doo Seok
Issue Date
2019-06-30
Citation
Second IT Revolution, and Dynamic Open Innovation; From Smart City, Autonomous Car, Intelligent Robot, and Block Chain to Sharing Economy
Type
Conference
Abstract
There are many models that represent a student knowledge and content that can be used for personalized instruction, analysis of learning and content analysis. The idea of knowledge representation lies in a high dimensional embedding to capture the dynamics of learning and testing. And knowledge tracing is a kind of sequence model that evaluate the student knowledge variation over time. So we can predict how a student will behave in the future. Recently, the results of using the artificial neural network approach such as RNN (Recurrent Neural Network) and deep reinforcement learning have a significant improvement in the prediction performance of knowledge tracing. In this paper, the author tries to evaluate a student knowledge not in the skills to specific problems but in the the degree of meta-cognitive skills. Using a large-scale data set collected in real-world classrooms, the author adopt the meta leaning machine models to successfully predict student learning outcomes and to improve personalized learning.
URI
http://hdl.handle.net/20.500.11750/14385
Publisher
SOItmC & Meijo University 2019
Related Researcher
  • Author Lee, Doo Seok  
  • Research Interests CAGD(Computer Aided Geometric Design); Optimization; Smart Education; Machine Learning
Files:
There are no files associated with this item.
Collection:
School of Undergraduate Studies2. Conference Papers


qrcode mendeley

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.

BROWSE