See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/356897159

Systematic Review for AI-based Language Learning Tools

Article in Journal of Digital Contents Society · November 2021

DOI: 10.9728/dcs.2021.22.11.1783

CITATION READS

1 34

2 authors:

Lauren J. Woo Heeyoul "Henry" Choi

Arizona State University Handong Global University
8 PUBLICATIONS 3 CITATIONS 76 PUBLICATIONS 799 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Heeyoul "Henry" Choi on 28 February 2022.

The user has requested enhancement of the downloaded file.

Systematic Review for AI-based Language Learning Tools
Jin Ha Woo1, Heeyoul Choi2

1
Arizona State University, Tempe, AZ
2
Handong Global University, Pohang, South Korea

Abstract
The Second Language Acquisition field has been significantly impacted by a greater emphasis on
individualized learning and rapid developments in artificial intelligence (AI). Although increasingly
adaptive language learning tools are being developed with the application of AI to the Computer
Assisted Language Learning field, there have been concerns regarding insufficient information and
teacher preparation. To effectively utilize these tools, teachers need an in-depth overview on recently
developed AI-based language learning tools. Therefore, this review synthesized information on AI tools
that were developed between 2017 and 2020. A majority of these tools utilized machine learning and
natural language processing, and were used to identify errors, provide feedback, and assess language
abilities. After using these tools, learners demonstrated gains in their language abilities and knowledge.
This review concludes by presenting pedagogical implications and emerging themes in the future
research of AI-based language learning tools.

Keyword: Artificial Intelligence, Computer Assisted Language Learning,, Educational Technology,

Language Learning

Preprint submitted to JDCS Oct. 29, 2021

Ⅰ. Introduction address the lack of evidence verifying the language
learning effectiveness of AI, efforts should be made to
The complex process of language learning involves acquire information on the pedagogical effects and
a variety of factors leading to numerous possible learner perceptions of AI-based language learning
outcomes. In particular, language learning is tools [5]. With this information, teachers can gain a
influenced by the degree of acculturation, amount of deeper awareness of available AI-based tools which
comprehensible input, attentiveness to L2 will enable them to facilitate the use of these tools
features/aspects, and opportunities for meaningful effectively and appropriately.
negotiation and production [1]. As people learn Presently, there is a lack of comprehensive reviews
additional languages, they make intentional efforts to on available AI-based language learning tools and the
develop their language knowledge and abilities. This pedagogical effects and learner perceptions of these
process involves understanding language components tools. Existing reviews related to AI in language
(e.g., grammar, vocabulary) that will help them to learning have focused on a specific type of AI-based
master the four core skills (reading, writing, listening, tool or the overall impact of AI on the future of
speaking) and related skill aspects (e.g., language education [4]-[5]. Thus, an in-depth
pronunciation) [2]. Over the past two decades, the overview is needed on recently developed AI-based
Complex Dynamic Systems theory and the multilingual language learning tools, including information on the
turn have led to a greater emphasis on individual types of utilized AI technology like NLP or machine
learner differences in language learning [3]. In learning (ML) and the targeted language skill areas
response to an emphasis on individualized learning, (e.g., reading, speaking) [7].
Intelligent Computer Assisted Language Learning In the next section, we describe our methodological
(ICALL) clearly emerged as a subfield of Computer approach, the research questions and the systematic
Assisted Language Learning (CALL) [4]. The ICALL review guidelines. Then, we present our findings based
subfield focuses on applying AI concepts, techniques, on our analysis of the relevant literature. Finally, we
algorithms, and technologies to CALL, especially conclude by discussing plans for future research.
natural language processing (NLP), user modeling,
expert systems, and intelligent tutoring systems.
Based on existing reviews related to AI in language
Ⅱ. Methods
learning, there has been a focus on developing
2-1 Research Aim and Study Design
tutoring systems, writing assistants, virtual reality
environments, chatbots, and other types of adaptive This review seeks to identify trends in the
learning systems/software [5]. The main intent of development of AI-based language learning tools and
these tools has been to generate personalized and provide detailed information on these tools. To
customizable learning experiences for the purposes of address the concerns regarding insufficient
optimizing language learning by increasing autonomy, information and teacher preparation identified in the
motivation, engagement, and effectiveness [5]-[6]. For existing reviews, we aimed to answer the following
instance, NLP-based tutoring systems are designed to questions:: What types of AI tools have been developed
provide tailored feedback, recommendations, and for various target language skill areas? How have
materials [4]. Recently, with the rapid development of these tools impacted language learning?
AI, these tools can meticulously adapt content in To answer these questions, we analyzed the
real-time to the learning pace, preferences, and needs contents of articles published in peer-reviewed
(e.g., cognitive, affective, social) of each user [4]-[5]. journals from 2017 to 2020. We decided to narrowly
Despite the immense potential of AI in language focus on this three-year period because of the recent
learning, there have been concerns regarding acceleration in mainstreaming AI into education. In
insufficient privacy, information, and teacher 2017, the global venture capital investment in AI
preparation. Foremost, as data collection is essential reached $1047 billion, and there has been a massive
to AI development, there is a need to reinforce privacy growth in AI development leading to the
policies and informed consent practices [4]. Also, to implementation of innovative techniques, algorithms,
approaches, and models [8]. search identified 415 results.
The review was conducted based on the Preferred For the screening process, we began by removing
Reporting Items for Systematic Reviews and results that were duplicated or could not be accessed
Meta-Analysis (PRISMA) guidelines, designed to through our institution system. Then, we examined the
transparently report on the purposes, methods, and relevancy of the title and abstract and removed results
findings of studies [9]. Following the PRISMA that did not mention utilizing AI for language learning.
guidelines, we went through the process of identifying After fully reviewing the remaining 62 results, we
and screening studies on AI-based language learning removed six studies that were non-empirical and three
tools, as shown in Figure 1. studies with insufficient information on the methods
Foremost, we identified studies using the Education and findings. Finally, 53 articles were included in this
Resources Information Center (ERIC), Scopus, and mixed-methods systematic review.
Web of Science database. These reputable databases
were selected because of their accessibility and wide 2-2 Data Analysis
coverage of academic journals related to education.
For the final sample of 53 articles, we developed a
To identify trends in the development of AI-based
spreadsheet template with categories quantitatively or
language learning tools, we conducted an initial
qualitatively describing various aspects of each article.
search using the following terms: “(artificial
To better understand the characteristics of the
intelligence) AND ((language learning) OR (second
articles, we compiled quantitative descriptions of the
language learning) OR (foreign language learning) OR
publication year, study design (methodology, location,
(EFL) OR (ESL))”. The search criteria were limited to
target language/language skill), and participants
peer-reviewed academic journals in English published
(grade, target language proficiency level). The target
from 2017 to 2020. As this initial search identified only
language skills covered in this review include language
39 results, we conducted a follow-up search using the
components (grammar, vocabulary), macro-skills
following specific terms related to AI and language
(listening, speaking, reading, writing), and a related
learning: “(ICALL) OR ((language learning) AND
skill aspect (pronunciation). In addition, to better
((natural language processing) OR (machine learning)
understand the types of tools that have been
OR (intelligent tutor) OR (bot) OR (virtual assistant)))
developed and the impact of these tools on language
OR ((artificial intelligence) AND ((language education)
learning, we compiled qualitative descriptions of the
OR (language development)))”. The search criteria
utilized AI technology (e.g., NLP, ML) and findings
were limited to peer-reviewed academic journals in
(pedagogical effects and learner perceptions).
English published from 2017 to 2020. The follow-up
To maintain consistency in the qualitative
descriptions, we reviewed 10 articles at a time and
took independent notes on the technology and
findings. Then, we met to discuss and compile our
notes. Through these discussions, we refined the
spreadsheet categories, adjusted the level of detail in
our notes, and worked towards identifying possible
trends in the data.
In examining the distribution of studies across
publication years (2017-2020), we found that many of
the studies (N = 35, 66%) were published in 2019 or
2020. Moreover, in examining the distribution of the
study design, we noticed that a significant number of
studies were quantitative or mixed (N = 50, 94%),
targeted the English language (N = 37, 69%), and
targeted the speaking and listening (N = 14, 26%),
writing (N = 11, 21%), or pronunciation (N = 11, 21%)
Fig. 1. PRISMA chart about the review phases [9].
 Fig. 6. Participants’ target language proficiency
Fig. 2.. Studies that took place in countries

levels. A detailed breakdown of the study design

(location, target language/skill) and participants
(grade, target language proficiency level) can be found
in Figure 2 to Figure 6.

Ⅲ. Results

3-1 Types of AI Tools Developed for Language Skill Areas

Fig. 3. Studies targeting various languages
In response to the question, ‘what types of AI tools
have been developed for various target language skill
areas?’, we provide an overview of the tools that have
been developed for each of the aforementioned skill
areas (speaking, listening, writing, pronunciation,
grammar, vocabulary, and reading) with the type of
tool (e.g., robots, mobile applications, and vitual
assistants) and AI technology.

1) Speaking and Listening

In the final sample, 14 studies (26%) identified a tool
targeting the speaking and listening skill. Four studies
Fig. 4. Targeted language skill area
explored the potential of using intelligent personal
assistants like Alexa by examining comprehensibility,
usability, and improvements in listening
comprehension, speaking proficiency, and willingness
to communicate [10]-[13]. Additionally, programmable
robots were used in group conversations ]14], and a
neural network (NN)-based dialogue system was used
for free conversation practice [15]. An NN-based
multimodal dialog system was also developed to
holistically assess spoken language in terms of
Fig. 5. Participants’ grade level
delivery, content, vocabulary, and grammar [16].

area. The results of the descriptive statistics also 2) Writing

revealed that a substantial number of the studies Eleven studies (21%) identified a tool targeting the
focused on university students (N = 32, 60%). Although writing skill area. The tools included machine
more studies focused on beginners (N = 10, 19%), there translators [17], software for free-form writing [18],
was a relatively even distribution across proficiency and a blended course with automated feedback on
self-correcting tasks [19]. There were also specialized was developed to identify pedagogical factors
systems focused on citations and referencing [20] and distinguishing high-achieving from low-achieving
classifying sentences into rhetoric categories [21]. readers to improve ESL reading instruction [35].
These tools incorporated the Latent Semantic
Analysis, random forests, support vector machines In summary, the AI tools identified in the 53 studies
(SVM), and Naive
̈ Bayes classifiers in ML. targeted diverse language skill areas and incorporated
many types of features, configurations, and
3) Pronunciation capabilities. For many of these tools, ML and NLP were
Eleven studies (21%) identified a tool targeting the incorporated into the configuration. While NLP
pronunciation skill area. In 7 of these studies, deep techniques (e.g., POS annotation, language modeling
learning algorithms [22] were used. Pronunciation and machine translation [36]) were used more often
diagnosis, training, and evaluation systems were for the writing, grammar, vocabulary, and reading
developed using the attention mechanism and various tools, NNs were used more often for the speaking,
types of NN (e.g., convolutional, long-short term listening, and pronunciation tools.
memory) [23]-[27]. For instance, a multimodal system
illustrating speech features [28] and an interactive tool 3-2 Impact of AI Tools on Language Learning
generating personalized voice models [29] have
In response to the question, ‘how have these tools
recently been developed.
impacted language learning?’, we provide an overview
4) Grammar of the pedagogical effects and learner perceptions of
Seven studies (13%) identified a tool targeting the the AI-based tools identified in the previous section.
grammar area. These tools included games, For each skill area, we summarize the intended
applications, immersive environments, and intelligent purposes of these tools as well as the changes in the
systems that utilized NN, ML, and NLP. For example, to language learning processes (e.g., experiences,
create customized study plans, NN modeling was used development of abilities/attitudes/ knowledge)
to predict grammatical challenges that learners may demonstrated by the learners and perceptions
encounter based on their first language [30]. For the self-reported by the learners after using these tools.
applications and systems, word segmentation, To enhance language learning processes, AI tools
syntactic parsing, and the finite state transducer in were used to identify errors, provide feedback, push
NLP were used to generate feedback [31]-[32]. resources, and assess/evaluate language abilities.
After using these tools, the learners demonstrated
5) Vocabulary gains in their language abilities, attitudes, knowledge,
Five studies (9%) identified a tool targeting the and use. In general, the learners perceived these tools
vocabulary area. These tools included systems, as effective, efficient, accurate, easy to use, and
platforms, robots, games, and mobile applications that useful/helpful for language learning. Overall, the
have been developed using ML (e.g., conditional learners reported having interesting, enjoyable, and
random field models) and NLP. For instance, in an satisfactory experiences with these tools.
ICALL platform, part-of-speech (POS) annotation and
syntactic parsing in NLP were used to visually enhance 1) Correcting Errors in Grammar
targeted vocabulary items by automatically generating AI-based grammar tools identified errors and
multiple-choice gaps [33]. One study (2%) identified a provided on-topic feedback [31]-[32]. By using these
mobile app with adaptive learning technology that tools, the learners were able to use English articles
targeted both the vocabulary and grammar area [34]. more accurately [37] and experience a greater sense of
immersion, presence, and realism while learning [38].
6) Reading Out of the 45 participants in [24], the 15 participants
The remaining four studies (8%) identified intelligent who used a digital game demonstrated significant
systems targeting the reading skill area. These gains on writing tasks eliciting the use of English
systems utilized ML to diagnose reading problems and articles (p = 0.000, Pre-test M = 61.6, Post-test M =
push appropriate resources. Additionally, an ML model 73.866, Delayed Post-test M = 84.800). The digital
game-based group significantly outperformed (p = as interesting, easy to use, useful, and helpful for
0.000) the participants in the cloze exercise group (N = language learning [48]-[49]. Furthermore, in a survey
15) and cloze exercise group with corrections (N = 15) administered by [50], the learners indicated
on the writing tasks. In regard to perceptions, the satisfaction with various aspects of a serious language
learners viewed these tools as effective, efficient, game (e.g., accessibility, skills acquisition, game
accurate, enjoyable, satisfactory, and easy to use. The mechanics, challenge/reward balance).
learners also noted that the tools adequately
represented their course materials and helped them 4) Improving Writing
achieve their language learning outcomes [32], [39]. AI-based writing tools identified errors, provided
feedback, assessed writing abilities, and facilitated
2) Assessing and Evaluating Conversations process-based academic writing [18], [20], [51]-[53].
For speaking and listening, AI tools were used to With these tools, the learners were able to reduce
assess speaking abilities, evaluate conversations, and plagiarism, increase editing/revising time, and correct
provide appropriate responses in open conversations rhetorical function, lexical, and grammatical errors
[15], [40]-[41]. With these tools, the learners became [17], [20]-[21], [54]. Following the use of a machine
more confident, willing, and less anxious about translator for a writing task involving multiple drafts,
speaking in English [13], [42]. The learners also the 34 participants in [17] demonstrated significant
demonstrated gains in listening and speaking in terms gains in their writing score (p = 0.000, initial draft M =
of pragmatics, cohesion, word concreteness, and use 3.76, final draft M = 4.56) and decreases in the number
of grammatical patterns [11], [14], [43]-[45]. Regarding of lexical (initial M = 5.97, final M = 3.82) and
perceptions, the learners indicated that the tools were grammatical errors (initial M = 15.67, final M = 9.82) (p
easy to use, authentic, comprehensible, and useful for = 0.000). After using a feedback system, the learners
language learning [10]-[12], [41], [43], [45]. The 29 also demonstrated significant improvements in their
participants in [10] experimented with Google Assistant essay drafts in terms of the organization, structure,
(GA) for an hour and then completed a 5-point coherence, supporting ideas, and conclusion [54].
Likert-scale questionnaire investigating the potential Furthermore, regarding perceptions, the learners
of using GA for language learning. Based on the items stated that these tools were effective, easy to use, and
with a mean score above a four, the participants felt useful/helpful for language learning [17], [20]-[21],
that GA could boost motivation to improve English [51], [55]. The learners also stated that these tools
listening abilities (M = 4.24) and speaking fluency (M = helped them to identify their writing
4.00), reduce stress when practicing English listening strengths/weaknesses and increase their writing
(M = 4.17) and speaking (M = 4.07), improve English knowledge through detailed comments [51].
listening comprehension abilities (M = 4.28), and
become an enjoyable hands-on tool to use (M = 4.07). 5) Improving Fluency in Pronunciation
For pronunciation, tools were used to detect
3) Suggesting Words mispronunciations and recognize speech for
For vocabulary, AI tools automatically detected diagnosis, assessment, and evaluation [23], [25]-[27],
Japanese expressions for the purpose of providing [56]. These tools helped the learners improve their
morphological analyses and example sentences [46]. fluency, comprehensibility, tone, and pronunciation
After using these tools, the learners demonstrated accuracy [24], [28]-[29], [57]. With regard to
gains in emotion word use and semantic knowledge of perceptions, the learners described these tools as
phrasal verbs [33], [47]. After a 3-week treatment with interesting, easy to use, and helpful for fluency,
an ML-based emotion synonym suggestion system, the intonation, and tone training [24], [29], [57].
33 participants in [47] demonstrated significant gains
(p < 0.01) on writing tasks evaluating emotion word use 6) Personalized Tutoring
with an average score increase of 1.77 (avg. pre-test AI-based reading tools classified learners, assessed
score = 3.41, highest possible score = 6). In regard to reading abilities, and pushed resources [35], [58]-[59].
perceptions, the learners generally viewed these tools For example, an adaptive learning system pushed
resources that corresponded to learner experiment with new technologies, they will gain the
characteristics (e.g., reading abilities, cognitive styles, knowledge and experiences needed to innovatively
learning objectives) [59]. After using an intelligent implement and facilitate student-centered technology
tutoring system, the learners demonstrated use.
improvements in essential academic reading skills, Lastly, learners have generally perceived many of
including main idea, text structure, and inference [58]. these tools as effective, interesting, easy to use, and
helpful. However, to determine the types of AI tools
To summarize, the AI tools identified in the 53 that work best for specific types of learners, teachers
studies have had a positive impact on language and researchers can administer open-ended surveys
learning and have been well-received by learners. and conduct interviews to better understand the
These tools identify errors, provide feedback, and reasoning behind their perceptions.
assess language abilities. Moreover, they have helped This review was limited to articles in peer-reviewed
learners to build on their language abilities and academic journals published in English from 2017 to
enhance their learning experiences. In particular, for 2020. Moreover, only three databases were used to
grammar and speaking/listening, the AI tools identify articles, and some relevant articles could not
positively impacted psychological factors such as be accessed through our institution system. This could
confidence, anxiety, immersion, and presence. By have been the reason why the studies were skewed
placing a greater emphasis on psychological factors, towards university students learning English in China
teachers can help learners to increase their or Japan.
comprehensible input and attentiveness to L2 Researchers plan to make systems more effective
features/aspects leading to greater gains [60]. and improve the accuracy of detection/recognition
capabilities. Additionally, researchers need to
Ⅳ. Discussions incorporate pedagogical knowledge by collaborating
with teachers and examining in-class use of these
Based on the reviewed studies, it is clear that tools. Researchers have also emphasized the need to
various AI tools targeting the speaking, listening, conduct studies that include diverse learners, examine
writing, pronunciation, grammar, vocabulary, and the long-term impact on learners, and verify
reading area have been developed. After using these contributions to language learning. Lastly, there are
tools, learners have demonstrated improvements in plans to expand their user bases by making the AI
their language skills/knowledge and perceived these tools available in languages other than English and
tools to be useful for language learning. adding more materials (e.g., feedback, exercises) for a
Since learners work directly with these AI tools, wider range of proficiency levels. Also, there is a need
learners and teachers need to become familiarized to monitor how recently developed AI tools including
with the fundamentals of commonly utilized AI large-scale deep learning models including BERT and
technologies. By bringing awareness to a wide array of GPT-3 will be applied to language learning in the near
available AI tools, teachers can guide learners to future.
select the most appropriate tools for their language
learning preferences and needs. For example, learners
References
who prioritize familiarity and accessibility can work
[1] L. Ortega, Understanding second language acquisition. New
with existing technologies such as Alexa and Google
York: Routledge, 2014
Translate. On the other hand, learners who prioritize
[2] R. L. Oxford, Teaching and researching language learning
collaborative learning can work with innovative tools
strategies: Self-regulation in context. New York: Taylor &
such as NAO, a programmable robot for group
Francis, 2016.
conversations.
[3] R. Ellis, “A short history of SLA: Where have we come from
Additionally, since there are tools targeting very
and where are we going?” Language Teaching, Vol. 54, No.
specific language abilities, teachers who are unfamiliar
2, pp. 190-205, 2021.
with AI can begin by experimenting for an activity
https://doi.org/10.1017/S0261444820000038
without a long-term commitment. As teachers
[4] J. Kannan, and P. Munday, “New trends in second language Implicit Learning in Joining-in-type Robot-assisted
learning and teaching through the lens of ICT, networked Language Learning System,” Int’l Journal of Emerging
learning, and artificial intelligence,” Círculo de Lingüística Technologies in Learning, Vol. 14, No. 2, 2019.
Aplicada a la Comunicación, 2018. https://doi.org/10.3991/ijet.v14i02.9212
https://doi.org/10.5209/CLAC.62495 [15] J. X. Huang, K. S. Lee, O. W. Kwon, and Y. K. Kim, “A
[5] S. Pokrivcakova, “Preparing teachers for the application of chatbot for a dialogue-based second language learning
AI-powered technologies in foreign language education,” system,” EUROCALL, pp. 151-156, 2017.
Journal of Language and Cultural Education, Vol. 7, No. 3, https://doi.org/10.14705/rpnet.2017.eurocall2017.705
pp. 135-153, 2019. https://doi.org/10.2478/jolace-2019-0025 [16] D. Suendermann-Oeft, V. Ramanarayanan, Z. Yu, Y. Qian,
[6] J. Woo and H. Choi, “Utilizing Machine Translation K. Evanini, P. Lange, X. Wang, and K. Zechner, “A
Systems to Generate Word Lists for Learning Vocabulary in multimodal dialog system for language assessment: Current
English,” Journal of Digital Contents Society, Vol. 22, No. state and future directions,” ETS Research Report Series,
1, pp. 71-75, 2021. https://doi.org/10.9728/dcs.2021.22.1.71 Vol. 2017, No. 1, pp. 1-7, 2017.
[7] J. Woo, “Oxford Learner’s Bookshelf: An e-book reader https://doi.org/10.1002/ets2.12149
application for extensive reading,” Univ. of Hawaii National [17] S. M. Lee, “The impact of using machine translation on
Foreign Language Resource Center, Vol. 30. No. 2, pp. EFL students’ writing,” CALL, Vol. 33, No. 3, pp. 157-175,
240-245, 2018. https://doi.org/10125/66926 2020. https://doi.org/10.1080/09588221.2018.1553186
[8] C. Guan, J. Mou, and Z. Jiang, “Artificial intelligence [18] R. Chacón-Beltrán, “Free-form writing: computerized
innovation in education: a twenty-year data-driven historical feedback for self-correction,” ELT Journal, Vol. 71, No. 2,
analysis,” Int’l J. of Innovation Studies, Vol. 4, No. 4, pp. pp. 141-149, 2017. https://doi.org/10.1093/elt/ccw064
134-147, 2020. https://doi.org/10.1016/j.ijis.2020.09.001 [19] C. Sarré, M. Grosbois, and C. Brudermann, “Fostering
[9] M. Page, J. McKenzie, P. Bossuyt, I. Boutron, T. Hoffmann, accuracy in L2 writing: impact of different types of
C. D. Mulrow, L. Shamseer, J. M. Tetzlaff, E. A. Akl, S. E. corrective feedback in an experimental blended learning
Brennan, R. Chou, J. Glanville, J. M. Grimshaw, A. EFL course,” CALL, 1-23, 2019.
Hrobjartsson, M. M. Lalu, T. Li, E. W. Loder, E. https://doi.org/10.1080/09588221.2019.1635164
Mayo-Wilson, S. McDonald, L. A. McGuinness, L. A. [20] W. C. Hsu, and G. Z. Liu, “Genre-based writing instruction
Stewart, J. Thomas, A. C. Tricco, V. A. Welch, P. Whiting, blended with an online writing tutorial system for the
and D. Moher, “The PRISMA 2020 statement: an updated development of academic writing,” Digital Scholarship in
guideline for reporting systematic reviews,” BMJ, 372, the Humanities, Vol. 34, No. 1, pp. 100-123, 2019.
2021. https://doi.org/10.31222/osf.io/v7gm2 https://doi.org/10.1093/llc/fqy021
[10] H. H. J. Chen, C. T. Y. Yang, and K. K. W. Lai, [21] M. A. Abbas, S. Hammad, G. J. Hwang, S. Khan, and S. M.
“Investigating college EFL learners’ perceptions toward the M. Gilani, “An assistive environment for EAL academic
use of Google Assistant for foreign language learning,” writing using formulaic sequences classification,”
Interactive Learning Environments, 1-16, 2020. Interactive Learning Environments, 1-15, 2020.
https://doi.org/10.1080/10494820.2020.1833043 https://doi.org/ 10.1080/10494820.2020.1789670
[11] G. Dizon, “Evaluating intelligent personal assistants for L2 [22] H. Choi, and Y. Min, “Intelligent Information System;
listening and speaking development,” Language Learning & Introduction to deep learning and major issues,” Korea
Technology, Vol. 24, No. 1, pp. 16-26, 2020. Information Processing Society Review, Vol. 22, No.1, pp.
https://doi.org/10125/44705 7-21, 2015.
[12] S. Moussalli, and W. Cardoso, “Intelligent personal [23] S. Akhtar, F. Hussain, F. R. Raja, M. Ehatisham-ul-haq, N.
assistants: can they understand and be understood by K. Baloch, F. Ishmanov, and Y. B. Zikria, “Improving
accented L2 learners?” CALL, Vol. 33, No. 8, pp. 865-890, mispronunciation detection of arabic words for non-native
2020. https://doi.org/10.1080/09588221.2019.1595664 learners using deep convolutional neural network features,”
[13] T. Y. Tai, and H. H. J. Chen, “The impact of Google Electronics, Vol. 9, No. 6, pp. 963, 2020.
Assistant on adolescent EFL learners’ willingness to https://doi.org/10.3390/electronics9060963
communicate,” Interactive Learning Environments, pp. [24] M. Kan, and A. Ito, “Language Cognition and
1-18, 2020. https://doi.org/10.1080/10494820.2020.1841801 Pronunciation Training Using Applications,” Future
[14] A. Khalifa, T. Kato, and S. Yamamoto, “Learning Effect of Internet, Vol. 12, No. 3, pp. 42, 2020.
 https://doi.org/10.3390/fi12030042 of a mobile gamification app on elementary students’
[25] K. Li, X. Qian, and H. Meng, “Mispronunciation detection Spanish achievement and self-efficacy,” CALL, Vol. 31, No.
and diagnosis in L2 english speech using multidistribution 1-2, pp. 72-89, 2018.
deep neural networks,” IEEE/ACM Trans. on Audio, Speech, https://doi.org/10.1080/09588221.2017.1382536
and Language Processing, Vol. 25, No. 1, pp. 193-207, [35] Y. Xiao, and J. Hu, “Assessment of optimal pedagogical
2017. https://doi.org/10.1109/TASLP.2016.2621675 factors for Canadian ESL learners’ reading literacy through
[26] J. Lin, Y. Gao, W. Zhang, L. Wei, Y. Xie, and J. Zhang, artificial intelligence algorithms,” Int. J. Engl. Linguist, Vol.
“Improving Pronunciation Erroneous Tendency Detection 9, pp. 1-14, 2019. https://doi.org/10.5539/ijel.v9n4p1
with Multi-Model Soft Targets,” Journal of Signal [36] H. Choi, “Understanding Neural Machine Translation,”
Processing Systems, 1-11, 2020. Communications of the KIISE, Vol. 37, No. 2, 2019.
https://doi.org/10.1007/s11265-019-01485-2 [37] C. W. Kao, “The effect of a digital game-based learning
[27] D. Mu, W. Sun, G. Xu, and W. Li, “Japanese Pronunciation task on the acquisition of the English Article System,”
Evaluation Based on DDNN,” IEEE Access, Vol. 8, pp. System, Vol. No. 95, pp. 102373, 2020.
218644-218657, 2020. https://doi.org/10.1016/j.system.2020.102373
https://doi.org/10.1109/ACCESS.2020.3041901 [38] Y. F. Wang, S. Petrina, and F. Feng, “VILLAGE- Virtual
[28] X. Liu, D. Wu, Y. Ye, M. Xu, J. Jiao, and W. Lin, Immersive Language Learning and Gaming Environment:
“Improving Accuracy in Imitating and Reading Aloud via Immersion and presence,” British Journal of Educational
Speech Visualization Technology,” Int’l Journal of Technology, Vol. 48, No. 2, pp. 431-450, 2017.
Emerging Technologies in Learning, Vol. 15, No. 8, 2020. https://doi.org/10.1111/bjet.12388
https://doi.org/10.3991/ijet.v15i08.11475 [39] C. Troussas, K. Chrysafiadi, and M. Virvou, “An intelligent
[29] S. Ding, C. Liberatore, S. Sonsaat, I. Lučić, A. Silpachai, G. adaptive fuzzy-based inference system for computer-assisted
Zhao, E. Chukharev-Hudilainen, J. Levis, and R. language learning,” Expert Systems with Applications, Vol.
Gutierrez-Osuna, “Golden speaker builder- An interactive 127, pp. 85-96, 2019.
tool for pronunciation training,” Speech Communication, https://doi.org/10.1016/j.eswa.2019.03.003
Vol. 115, pp. 51-66, 2019. [40] Y. Lee, and J. Cho, “Design of a Foreign Language
https://doi.org/10.1016/j.specom.2019.10.005 Conversation Learning System Using Machine Learning,”
[30] E. Kartal, “The Artificial Neural Network Modeling of Int’l Journal of Innovation, Creativity and Change, Vol. 13,
Language Learning Challenges of French-Speaking Students No. 4, pp. 235-247, 2020.
Learning Turkish as a Foreign Language: The Case of [41] K. C. Li, M. Chang, K. H. Wu, “Developing a Task-Based
France,” Educational Sciences: Theory and Practice, Vol. Dialogue System for English Language Learning,”
19, No. 1, 55-77. 2017. Education Sciences, Vol. 10. No. 11, pp. 306, 2020.
https://doi.org/10.12738/estp.2019.1.0264 https://doi.org/10.3390/educsci10110306
[31] H. Ai, “Providing graduated corrective feedback in an [42] E. Ayedoun, Y. Hayashi, and K. Seta, “Adding
intelligent computer-assisted language learning communicative and affective strategies to an embodied
environment,” ReCALL, Vol. 29, No. 3, pp. 313-334, 2017. conversational agent to enhance second language learners’
https://doi.org/10.1017/S095834401700012X willingness to communicate,” Int’l Journal of Artificial
[32] M. Bontogon, A. Arppe, L. Antonsen, D. Thunder, and J. Intelligence in Education, Vol. 29, No. 1, pp. 29-57, 2019.
Lachler, “Intelligent computer assisted language learning https://doi.org/10.1007/s40593-018-0171-6
(ICALL) for nêhiyawêwin: an in-depth user-experience [43] C. M. Forsyth, C. Luce, D. Zapata-Rivera, G. T. Jackson,
evaluation,” Canadian Modern Language Review, Vol. 74, K. Evanini, and Y. So, “Evaluating English language
No. 3, pp. 337-362, 2018. https://doi.org/10.3138/cmlr.4054 learners’ conversations: Man vs. Machine,” CALL, Vol. 32,
[33] S. Ruiz, P. Rebuschat, and D. Meurers, “The effects of No. 4, pp. 398-417, 2019.
working memory and declarative memory on instructed https://doi.org/10.1080/09588221.2018.1517126
second language vocabulary learning: Insights from [44] K. Scholz, “Encouraging free play: Extramural digital
intelligent CALL,” Language Teaching Research, Vol. 25, game-based language learning as a complex adaptive
No. 4, pp. 510-539, 2021. system,” CALICO Journal, Vol. 34, No. 1, pp. 39-57, 2017.
https://doi.org/10.1177/1362168819872859 https://doi.org/10.1558/cj.29527
[34] J. R. Rachels, and A. J. Rockinson-Szapkiw, “The effects [45] T. Sydorenko, P. Daurio and L. S. Thorne, “Refining
 pragmatically-appropriate oral communication via and writing course,” Library Hi Tech, 2020.
computer-simulated conversations,” CALL, Vol. 31, No. 1-2, https://doi.org/10.1108/LHT-05-2020-0113
pp. 157-180, 2018. [56] S. Detey, L. Fontan, M. Le Coz, and S. Jmel,
https://doi.org/10.1080/09588221.2017.1394326 “Computer-assisted assessment of phonetic fluency in a
[46] J. Liu, H. Shindo, and Y. Matsumoto, “Development of a second language: a longitudinal study of Japanese learners
computer-assisted Japanese functional expression learning of French,” Speech Communication, Vol. 125, pp. 69-79,
system for Chinese-speaking learners,” Educational 2020. https://doi.org/10.1016/j.specom.2020.10.001
Technology Research and Development, Vol. 67, No. 5, pp. [57] X. Peng, H. Chen, L. Wang, and H. Wang, “Evaluating a
1307-1331, 2019. 3-D virtual talking head on pronunciation learning,” Int’l
https://doi.org/10.1007/s11423-019-09669-0 Journal of Human-Computer Studies, Vol. 109, pp. 26-40,
[47] M. H. Chen, W. F. Chen, and L. W. Ku, “Application of 2018. https://doi.org/10.1016/j.ijhcs.2017.08.001
sentiment analysis to language learning,” IEEE Access, Vol. [58] C. Kaoropthai, O. Natakuatoong, N. Cooharojananone, “An
6, pp. 24433-24442, 2018. intelligent diagnostic framework: A scaffolding tool to
https://doi.org/10.1109/ACCESS.2018.2832137 resolve academic reading problems of Thai first-year
[48] E. Chew, and X. N. Chua, “Robotic Chinese language university students,” Computers & Education, Vol. 128, pp.
tutor: personalising progress assessment and feedback or 132-144, 2019.
taking over your job?” On the Horizon. Vol. 28 No. 3, pp. https://doi.org/10.1016/j.compedu.2018.09.001
113-124. 2020. https://doi.org/10.1108/OTH-04-2020-0015 [59] J. Zhang, M. Xie, and B. Wen, “Adaptive Pushing of
[49] H. Liu, L. Wang, and M. J. Koehler, “Exploring the Learning Resources in Fragmented English Reading,” Int’l
intention-behavior gap in the technology acceptance model: Journal of Performability Engineering, Vol. 15, No. 3, pp.
A mixed-methods study in the context of foreign-language 884, 2019. https://doi.org/10.23940/ijpe.19.03.p17.884894
teaching in China,” British Journal of Educational [60] H. Ni, “The Effects of Affective Factors in SLA and
Technology, Vol. 50, No. 5, pp. 2536-2556, 2019. Pedagogical Implications,” Theory and Practice in
https://doi.org/10.1111/bjet.12824 Language Studies, Vol. 2, No. 7, pp. 1508-1513, 2012.
[50] Y. Tazouti, S. Boulaknadel, and Y. Fakhri, “ImALeG: A https://doi.org/10.4304/tpls.2.7.1508-1513
Serious Game for Amazigh Language Learning,” Int’l
Journal of Emerging Technologies in Learning, Vol. 14, No.
18, pp. 28-38, 2019.
https://doi.org/10.3991/ijet.v14i18.10854
[51] G. Cheng, “Towards an automatic classification system for
supporting the development of critical reflective skills in L2
learning,” Australasian Journal of Educational Technology,
Vol. 33, No. 4, 2017. https://doi.org/10.14742/ajet.3029
[52] C. Harvey-Scholes, “Computer-assisted detection of 90%
of EFL student errors,” CALL, Vol. 31, No. 1-2, pp.
144-156, 2018.
https://doi.org/10.1080/09588221.2017.1392322
[53] H. Yannakoudakis, Ø. E. Andersen, A. Geranpayeh, T.
Briscoe, and D. Nicholls, “Developing an automated writing
placement system for ESL learners,” Applied Measurement
in Education, Vol. 31, No. 3, pp. 251-267, 2018.
https://doi.org/10.1080/08957347.2018.1464447
[54] M. Liu, Y. Li, W. Xu, and L. Liu, “Automated essay
feedback generation and its impact on revision,” IEEE Tran.
on Learning Technologies, Vol. 10, No. 4, pp. 502-513,
2017. https://doi.org/10.1109/TLT.2016.2612659
[55] Z. Wang, “Computer-assisted EFL writing and evaluations
based on artificial intelligence: a case from a college reading

View publication stats