東華大學圖書館 |

語系: 繁體中文

說明(常見問題)

回圖書館首頁

手機版館藏查詢

登入

回首頁

切換: 標籤 | MARC模式 | ISBD

Understanding and Supporting Vocabul...

Nam, Sungjin.

FindBook

Google Book

Amazon

博客來

Understanding and Supporting Vocabulary Learners Via Machine Learning on Behavioral and Linguistic Data.

紀錄類型:	書目-電子資源 : Monograph/item
正題名/作者:	Understanding and Supporting Vocabulary Learners Via Machine Learning on Behavioral and Linguistic Data./
作者:	Nam, Sungjin.
出版者:	Ann Arbor : ProQuest Dissertations & Theses, : 2020,
面頁冊數:	254 p.
附註:	Source: Dissertations Abstracts International, Volume: 82-07, Section: B.
Contained By:	Dissertations Abstracts International82-07B.
標題:	Curriculum development. -
電子資源:	https://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=28240166
ISBN:	9798684623875

Understanding and Supporting Vocabulary Learners Via Machine Learning on Behavioral and Linguistic Data.
Nam, Sungjin.

Understanding and Supporting Vocabulary Learners Via Machine Learning on Behavioral and Linguistic Data. - Ann Arbor : ProQuest Dissertations & Theses, 2020 - 254 p.

Source: Dissertations Abstracts International, Volume: 82-07, Section: B.

Thesis (Ph.D.)--University of Michigan, 2020.

This item must not be sold to any third party vendors.

This dissertation presents various machine learning applications for predicting different cognitive states of students while they are using a vocabulary tutoring system, DSCoVAR. We conduct four studies, each of which includes a comprehensive analysis of behavioral and linguistic data and provides data-driven evidence for designing personalized features for the system.The first study presents how behavioral and linguistic interactions from the vocabulary tutoring system can be used to predict students' off-task states. The study identifies which predictive features from interaction signals are more important and examines different types of off-task behaviors. The second study investigates how to automatically evaluate students' partial word knowledge from open-ended responses to definition questions. We present a technique that augments modern word-embedding techniques with a classic semantic differential scaling method from cognitive psychology. We then use this interpretable semantic scale method for predicting students' short- and long-term learning. The third and fourth studies show how to develop a model that can generate more efficient training curricula for both human and machine vocabulary learners. The third study illustrates a deep-learning model to score sentences for a contextual vocabulary learning curriculum. We use pre-trained language models, such as ELMo or BERT, and an additional attention layer to capture how the context words are less or more important with respect to the meaning of the target word. The fourth study examines how the contextual informativeness model, originally designed to develop curricula for human vocabulary learning, can also be used for developing curricula for various word embedding models. We identify sentences predicted as low informative for human learners are also less helpful for machine learning algorithms.Having a rich understanding of user behaviors, responses, and learning stimuli is imperative to develop an intelligent online system. Our studies demonstrate interpretable methods with cross-disciplinary approaches to understand various cognitive states of students during learning. The analysis results provide data-driven evidence for designing personalized features that can maximize learning outcomes. Datasets we collected from the studies will be shared publicly to promote future studies related to online tutoring systems. And these findings can also be applied to represent different user states observed in other online systems. In the future, we believe our findings can help to implement a more personalized vocabulary learning system, to develop a system that uses non-English texts or different types of inputs, and to investigate how the machine learning outputs interact with students.

ISBN: 9798684623875Subjects--Topical Terms:

684418
Curriculum development.
Subjects--Index Terms:

Intelligent tutoring system

Understanding and Supporting Vocabulary Learners Via Machine Learning on Behavioral and Linguistic Data.
LDR:04228nmm a2200457 4500 001 2278737
005 20210712062255.5
008 220723s2020 ||||||||||||||||| ||eng d
020 $a 9798684623875
035 $a (MiAaPQ)AAI28240166
035 $a (MiAaPQ)umichrackham003345
035 $a AAI28240166
040 $a MiAaPQ $c MiAaPQ
100 1 $a Nam, Sungjin. $3 3557130
245 1 0 $a Understanding and Supporting Vocabulary Learners Via Machine Learning on Behavioral and Linguistic Data.
260 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2020
300 $a 254 p.
500 $a Source: Dissertations Abstracts International, Volume: 82-07, Section: B.
500 $a Advisor: Collins-Thompson, Kevyn.
502 $a Thesis (Ph.D.)--University of Michigan, 2020.
506 $a This item must not be sold to any third party vendors.
506 $a This item must not be added to any third party search indexes.
520 $a This dissertation presents various machine learning applications for predicting different cognitive states of students while they are using a vocabulary tutoring system, DSCoVAR. We conduct four studies, each of which includes a comprehensive analysis of behavioral and linguistic data and provides data-driven evidence for designing personalized features for the system.The first study presents how behavioral and linguistic interactions from the vocabulary tutoring system can be used to predict students' off-task states. The study identifies which predictive features from interaction signals are more important and examines different types of off-task behaviors. The second study investigates how to automatically evaluate students' partial word knowledge from open-ended responses to definition questions. We present a technique that augments modern word-embedding techniques with a classic semantic differential scaling method from cognitive psychology. We then use this interpretable semantic scale method for predicting students' short- and long-term learning. The third and fourth studies show how to develop a model that can generate more efficient training curricula for both human and machine vocabulary learners. The third study illustrates a deep-learning model to score sentences for a contextual vocabulary learning curriculum. We use pre-trained language models, such as ELMo or BERT, and an additional attention layer to capture how the context words are less or more important with respect to the meaning of the target word. The fourth study examines how the contextual informativeness model, originally designed to develop curricula for human vocabulary learning, can also be used for developing curricula for various word embedding models. We identify sentences predicted as low informative for human learners are also less helpful for machine learning algorithms.Having a rich understanding of user behaviors, responses, and learning stimuli is imperative to develop an intelligent online system. Our studies demonstrate interpretable methods with cross-disciplinary approaches to understand various cognitive states of students during learning. The analysis results provide data-driven evidence for designing personalized features that can maximize learning outcomes. Datasets we collected from the studies will be shared publicly to promote future studies related to online tutoring systems. And these findings can also be applied to represent different user states observed in other online systems. In the future, we believe our findings can help to implement a more personalized vocabulary learning system, to develop a system that uses non-English texts or different types of inputs, and to investigate how the machine learning outputs interact with students.
590 $a School code: 0127.
650 4 $a Curriculum development. $3 684418
650 4 $a Educational administration. $3 2122799
650 4 $a Computer science. $3 523869
650 4 $a Artificial intelligence. $3 516317
650 4 $a Educational tests & measurements. $3 3168483
650 4 $a Educational technology. $3 517670
650 4 $a Linguistics. $3 524476
650 4 $a Behavioral sciences. $3 529833
650 4 $a Information science. $3 554358
653 $a Intelligent tutoring system
653 $a Machine learning
653 $a Natural language processing
653 $a Vocabulary tutoring
690 $a 0984
690 $a 0723
690 $a 0288
690 $a 0710
690 $a 0290
690 $a 0514
690 $a 0602
690 $a 0800
690 $a 0727
710 2 $a University of Michigan. $b Information. $3 2100858
773 0 $t Dissertations Abstracts International $g 82-07B.
790 $a 0127
791 $a Ph.D.
792 $a 2020
793 $a English
856 4 0 $u https://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=28240166