Intelligent Tutoring Systems (ITS) are computer programs that model learners' psychological states to provide individualized instruction. They have been developed for diverse subject areas (e.g., ...algebra, medicine, law, reading) to help learners acquire domain-specific, cognitive and metacognitive knowledge. A meta-analysis was conducted on research that compared the outcomes from students learning from ITS to those learning from non-ITS learning environments. The meta-analysis examined how effect sizes varied with type of ITS, type of comparison treatment received by learners, type of learning outcome, whether knowledge to be learned was procedural or declarative, and other factors. After a search of major bibliographic databases, 107 effect sizes involving 14,321 participants were extracted and analyzed. The use of ITS was associated with greater achievement in comparison with teacher-led, large-group instruction (g = .42), non-ITS computer-based instruction (g = .57), and textbooks or workbooks (g = .35). There was no significant difference between learning from ITS and learning from individualized human tutoring (g = -.11) or small-group instruction (g = .05). Significant, positive mean effect sizes were found regardless of whether the ITS was used as the principal means of instruction, a supplement to teacher-led instruction, an integral component of teacher-led instruction, or an aid to homework. Significant, positive effect sizes were found at all levels of education, in almost all subject domains evaluated, and whether or not the ITS provided feedback or modeled student misconceptions. The claim that ITS are relatively effective tools for learning is consistent with our analysis of potential publication bias.
Knowledge Tracing (KT) aims to trace the student’s state of evolutionary mastery for a particular knowledge or concept based on the student’s historical learning interactions with the corresponding ...exercises. Taking the “exercise-to-concept” relationships as input, several existing methods have been developed to trace and model students’ mastery states. However, these studies face two major shortcomings in KT: 1) they only consider “exercise-to-concept” relationships; 2) the multi-hot embeddings lack interpretability. In order to address the above issues, we propose a Joint graph convolutional network based deep Knowledge Tracing (JKT) framework to model the multi-dimensional relationships of “exercise-to-exercise”, and “concept-to-concept” into graph and fuse them with “exercise-to-concept” relationships. In JKT, it is not only possible to establish connections between exercises under cross-concepts, but also to help capture high-level semantic information and increase the model’s interpretability. In addition, sufficient experiments conducted on four real-world datasets have demonstrated that JKT performs better than the other baseline models. We further illustrate a case study to demonstrate its interpretability for learning analysis
Recent increase in the availability of learning data has given educational data mining an importance and momentum, in order to better understand and optimize the learning process and environments in ...which it occurs. The aim of this paper is to provide a comprehensive analysis and comparison of state of the art supervised machine learning techniques applied for solving the task of student exam performance prediction, i.e. discovering students at a “high risk” of dropping out from the course, and predicting their future achievements, such as for instance, the final exam scores. For both classification and regression tasks, the overall highest precision was obtained with artificial neural networks by feeding the student engagement data and past performance data, while the usage of demographic data did not show significant influence on the precision of predictions. To exploit the full potential of the student exam performance prediction, it was concluded that adequate data acquisition functionalities and the student interaction with the learning environment is a prerequisite to ensure sufficient amount of data for analysis.
•Overview of state of the art machine learning techniques in the context of student performance prediction.•Comparison of performance trends and computational requirements of analysed machine learning techniques.•Identification of optimal input data sets for optimisation of particular machine learning technique.•Artificial Neural Networks showed overall best results for solving student performance prediction tasks.
With the rapid growth of technology, computer learning has become increasingly integrated with artificial intelligence techniques in order to develop more personalized educational systems. These ...systems are known as Intelligent Tutoring systems (ITSs). This paper focused on the variant characteristics of ITSs developed across different educational fields. The original studies from 2007 to 2017 were extracted from the PubMed, ProQuest, Scopus, Google scholar, Embase, Cochrane, and Web of Science databases. Finally, 53 papers were included in the study based on inclusion criteria. The educational fields in the ITSs were mainly computer sciences (37.73%). Action-condition rule-based reasoning, data mining, and Bayesian network with 33.96%, 22.64%, and 20.75% frequency respectively, were the most frequent artificial intelligent techniques applied in the ITSs. These techniques enable ITSs to deliver adaptive guidance and instruction, evaluate learners, define and update the learner's model, and classify or cluster learners. Specifically, the performance of the system, learner's performance, and experiences were used for evaluation of ITSs. Most ITSs were designed for web user interfaces. Although these systems could facilitate reasoning in the learning process, these systems have rarely been applied in experimental courses including problem-solving, decision-making in physics, chemistry, and clinical fields. Due to the important role of a cell phone in facilitating personalized learning and given the low rate of using mobile-based ITSs, this study has recommended the development and evaluation of mobile-based ITSs.
The goal of Knowledge Tracing (KT) is to estimate how well students have mastered a concept based on their historical learning of related exercises. The benefit of knowledge tracing is that students’ ...learning plans can be better organised and adjusted, and interventions can be made when necessary. With the recent rise of deep learning, Deep Knowledge Tracing (DKT) has utilised Recurrent Neural Networks (RNNs) to accomplish this task with some success. Other works have attempted to introduce Graph Neural Networks (GNNs) and redefine the task accordingly to achieve significant improvements. However, these efforts suffer from at least one of the following drawbacks: (1) they pay too much attention to details of the nodes rather than to high-level semantic information; (2) they struggle to effectively establish spatial associations and complex structures of the nodes; and (3) they represent either concepts or exercises only, without integrating them. Inspired by recent advances in self-supervised learning, we propose a Bi-Graph Contrastive Learning based Knowledge Tracing (Bi-CLKT) to address these limitations. Specifically, we design a two-layer comparative learning scheme based on an “exercise-to-exercise” (E2E) relational subgraph. It involves node-level contrastive learning of subgraphs to obtain discriminative representations of exercises, and graph-level contrastive learning to obtain discriminative representations of concepts. Moreover, we designed a joint contrastive loss to obtain better representations and hence better prediction performance. Also, we explored two different variants, using RNN and memory-augmented neural networks as the prediction layer for comparison to obtain better representations of exercises and concepts respectively. Extensive experiments on four real-world datasets show that the proposed Bi-CLKT and its variants outperform other baseline models.
“Knowledge tracing (KT)” is an emerging and popular research topic in the field of online education that seeks to assess students’ mastery of a concept based on their historical learning of relevant ...exercises on an online education system in order to make the most accurate prediction of student performance. Since there have been a large number of KT models, we attempt to systematically investigate, compare and discuss different aspects of KT models to find out the differences between these models in order to better assist researchers in this field. The findings of this study have made substantial contributions to the progress of online education, which is especially relevant in light of the current global pandemic. As a result of the current expansion of deep learning methods over the last decade, researchers have been tempted to include deep learning strategies into KT research with astounding results. In this paper, we evaluate current research on deep learning-based KT in the main categories listed below. In particular, we explore (1) a granular categorisation of the technological solutions presented by the mainstream Deep Learning-based KT Models. (2) a detailed analysis of techniques to KT, with a special emphasis on Deep Learning-based KT Models. (3) an analysis of the technological solutions and major improvement presented by Deep Learning-based KT models. In conclusion, we discuss possible future research directions in the field of Deep Learning-based KT.
In the present study, we developed a chatbot that helps teachers to deliver writing instructions. By working with the chatbot, the post-secondary writers developed a thesis statement for their ...argumentative essay outlines, and the chatbot helped the writers to refine their peer review feedback. We conducted a preliminary analysis of the effect of a chatbot on these writers' writing achievement. We also collected several student testimonials about their chatbot experiences. Several important pedagogical and research implications for chatbot-guided writing instructions and the use of learning technology have been addressed.
The goal of this study was to investigate 65 students' evidence scores of emotions while they engaged in cognitive and metacognitive self-regulated learning processes as they learned about the ...circulatory system with MetaTutor, a hypermedia-based intelligent tutoring system. We coded for the accuracy of detecting students’ cognitive and metacognitive processes, and examined how the computed scores related to mean evidence scores of emotions and overall learning. Results indicated that mean evidence score of surprise negatively predicted the accuracy of making a metacognitive judgment, and mean evidence score of frustration positively predicted the accuracy of taking notes, a cognitive learning strategy. These results have implications for understanding the beneficial role of negative emotions during learning with advanced learning technologies. Future directions include providing students with feedback about the benefits of both positive and negative emotions during learning and how to regulate specific emotions to ensure the most effective learning experience with advanced learning technologies.
•Mean evidence score of contempt was significantly lower than joy or anger.•Mean evidence score of anger was significantly higher than frustration.•Emotions correlated with accuracy scores, but not proportional learning gain.•Surprise negatively predicted accuracy score of feeling of knowing.•Frustration positively predicted accuracy score of notes.
Artificial Intelligence (AI) plays an increasingly important role in language education; however, the trends, research issues, and applications of AI in language learning remain largely ...under-investigated. Accordingly, the present paper, using bibliometric analysis, investigates these issues via a review of 516 papers published between 2000 and 2019, focusing on how AI was integrated into language education. Findings revealed that the frequency of studies on AI-enhanced language education increased over the period. The USA and Arizona State University were the most active country and institution, respectively. The 10 most popular topics were: (1) automated writing evaluation; (2) intelligent tutoring systems (ITS) for reading and writing; (3) automated error detection; (4) computer-mediated communication; (5) personalized systems for language learning; (6) natural language and vocabulary learning; (7) web resources and web-based systems for language learning; (8) ITS for writing in English for specific purposes; (9) intelligent tutoring and assessment systems for pronunciation and speech training; and (10) affective states and emotions. The results also indicated that AI was frequently used to assist students in learning writing, reading, vocabulary, grammar, speaking, and listening. Natural language processing, automated speech recognition, and learner profiling were commonly applied to develop automated writing evaluation, personalized learning, and intelligent tutoring systems.
Pedagogical agents are typically designed to take on a single role: either as a tutor who guides and instructs the student, or as a tutee that learns from the student to reinforce what he/she knows. ...While both agent-role paradigms have been shown to promote student learning, we hypothesize that there will be heightened benefit with respect to students’ learning and emotional engagement if the agent engages children in a more peer-like way — adaptively switching between tutor/tutee roles. In this work, we present a novel active role-switching (ARS) policy trained using reinforcement learning, in which the agent is rewarded for adapting its tutor or tutee behavior to the child’s knowledge mastery level. To investigate how the three different child–agent interaction paradigms (tutee, tutor, and peer agents) impact children’s learning and affective engagement, we designed a randomized controlled between-subject experiment. Fifty-nine children aged 5–7 years old from a local public school participated in a collaborative word-learning activity with one of the three agent-role paradigms. Our analysis revealed that children’s vocabulary acquisition benefited from the robot tutor’s instruction and knowledge demonstration, whereas children exhibited slightly greater affect on their faces when the robot behaves as a tutee of the child. This synergistic effect between tutor and tutee roles suggests why our adaptive peer-like agent brought the most benefit to children’s vocabulary learning and affective engagement, as compared to an agent that interacts only as a tutor or tutee for the child. This work sheds light on how fixed role (tutor/tutee) and adaptive role (peer) agents support children’s cognitive and emotional needs as they play and learn. It also contributes to an important new dimension of designing educational agents — actively adapting roles based on the student’s engagement and learning needs.
•How agent roles (tutor/tutee/peer) impact children’s learning and affect was compared.•The peer agent’s role-switching mechanism was built using reinforcement learning.•The synergistic effect between tutor and tutee roles was observed in the peer agent.•The peer agent brought the most benefit to children’s learning and affect.•The guidance of designing educational agents for young children was discussed.