=Paper=
{{Paper
|id=Vol-3667/GenAILA-paper1
|storemode=property
|title=3DG: A Framework for Using Generative AI for Handling Sparse Learner Performance Data From Intelligent Tutoring Systems
|pdfUrl=https://ceur-ws.org/Vol-3667/GenAILA-paper1.pdf
|volume=Vol-3667
|authors=Liang Zhang,Jionghao Lin,Conrad Borchers,Meng Cao,Xiangen Hu
|dblpUrl=https://dblp.org/rec/conf/lak/ZhangLBCH24
}}
==3DG: A Framework for Using Generative AI for Handling Sparse Learner Performance Data From Intelligent Tutoring Systems ==
https://ceur-ws.org/Vol-3667/GenAILA-paper1.pdf
3DG: A Framework for Using Generative AI for
Handling Sparse Learner Performance Data From
Intelligent Tutoring Systems⋆
Liang Zhang1,2 , Jionghao Lin3,* , Conrad Borchers3 , Meng Cao4 and Xiangen Hu1,2,4
1
Institute for Intelligent Systems, University of Memphis, Memphis, TN 38152, USA
2
Department of Electrical and Computer Engineering, University of Memphis, Memphis, TN 38152, USA
3
Human-Computer Interaction Institute, Carnegie Mellon University, Pittsburgh, PA, 15213, USA
4
Department of Psychology, University of Memphis, Memphis, TN, 38152, USA
Abstract
Learning performance data (e.g., quiz scores and attempts) is significant for understanding learner
engagement and knowledge mastery level. However, the learning performance data collected from
Intelligent Tutoring Systems (ITSs) often suffers from sparsity, impacting the accuracy of learner modeling
and knowledge assessments. To address this, we introduce the 3DG framework (3-Dimensional tensor for
Densification and Generation), a novel approach combining tensor factorization with advanced generative
models, including Generative Adversarial Network (GAN) and Generative Pre-trained Transformer (GPT),
for enhanced data imputation and augmentation. The framework operates by first representing the data
as a three-dimensional tensor, capturing dimensions of learners, questions, and attempts. It then densifies
the data through tensor factorization and augments it using Generative AI models, tailored to individual
learning patterns identified via clustering. Applied to data from an AutoTutor lesson by the Center for
the Study of Adult Literacy (CSAL), the 3DG framework effectively generated scalable, personalized
simulations of learning performance. Comparative analysis revealed GAN’s superior reliability over
GPT-4 in this context, underscoring its potential in addressing data sparsity challenges in ITSs and
contributing to the advancement of personalized educational technology.
Keywords
Learning Performance Data, Data Sparsity, Intelligent Tutoring System, Generative Model, Generative
Adversarial Network, Generative Pre-trained Transformer
1. Introduction
Intelligent Tutoring System (ITS) is a prototype of computer system designed to offer personal-
ized and adaptive instructions through tracing and analyzing learning performance data such as
quiz scores and question attempts [1, 2, 3, 4]. However, during the interaction between learners
and ITS, learning performance data often exhibits data sparsity due to unexplored questions, in-
sufficient attempts to master knowledge, and lacking variability in learning patterns [5, 6, 7, 8, 9].
Joint Proceedings of LAK 2024 Workshops, co-located with 14th International Conference on Learning Analytics and
Knowledge (LAK 2024), Kyoto, Japan, March 18-22, 2024.
*
Corresponding author.
$ lzhang13@memphis.edu (L. Zhang); jionghao@cmu.edu (J. Lin); cborcher@cs.cmu.edu (C. Borchers);
mcao@memphis.edu (M. Cao); xhu@memphis.edu (X. Hu)
0009-0002-0017-2569 (L. Zhang); 0000-0003-3320-3907 (J. Lin); 0000-0003-3437-8979 (C. Borchers);
0000-0002-1286-2885 (M. Cao); 0000-0001-9045-4070 (X. Hu)
© 2024 Copyright for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).
CEUR
ceur-ws.org
Workshop ISSN 1613-0073
Proceedings
Data sparsity can lead to biased analysis and modeling of learning data. This is particularly
evident in the “Learner Model” component of ITS, which is crucial for tracking learning and
predicting performance of individual learners [10, 11, 12]. Specifically, sparse performance data
can lead to skewed or overfitted Knowledge Tracing models in “Learner Model”, which impedes
accurately capturing learner knowledge states and may result in misleading predictions of
learning performance [6, 9, 13, 14]. The scarcity of learning performance data significantly
hampers the development of ITSs, particularly in cases where learners have not sufficiently
engaged with certain instructional scenarios [15, 16].
Tackling data sparsity for ITSs presents a practical yet challenging research area. Informed
by the machine learning literature [17, 18, 19, 20], the issue of data sparsity can be addressed by
two principal ways: data imputation and data augmentation. Firstly, data imputation focuses
on filling the gaps in missing data to ensure a comprehensive dataset [6, 8, 21]. Secondly, data
augmentation aims to enrich and expand datasets where there are insufficient learning patterns,
thus ensuring robustness in analysis, modeling, and even potential testing tasks for ITSs [9, 22].
Currently, limited efforts have been made in the field of ITSs to systematically address these
data sparsity issues in learning performance data [5, 6, 21]. Driven by this, we propose the 3DG
(3-dimensions, Densification, and Generation) simulation framework, a systematic approach
leveraging generative models to handle sparse learning performance from ITS.
The 3DG framework derived from its three core phases. In the first phase, a 3-dimensional
tensor is constructed to represent learning performance data, with dimensions corresponding
to learners, questions, and attempts. The second phase focuses on densifying the sparse tensor
by tensor factorization. The third phase entails the generation of learning performance data
based on generative models, tailored to the individual learning patterns of learners. The 3DG
framework integrates the multidimensional learner model with generative models to facilitate
scalable simulation sampling for individual learning patterns. The multidimensional learner
model in our framework is derived from the Tensor Factorization method, a widely-used
approach in predicting learner performance in many studies [21, 23, 24, 25]. Initially, learning
performance values are represented in a three-dimensional tensor encompassing dimensions
of learners, questions, and attempts. Specifically, learning performance indicators, such as
binary responses from learners at problem-solving step attempts (with correct answers denoted
as 1 and incorrect as 0), form the tensor entries, and they are arranged sequentially along
the question queue in the learning process and sorted by attempts in ascending order. This
constructed tensor exhibits data sparsity. Our study aims to perform data imputation and
augmentation on the sparse tensor. Mathematically, the tensor factorization method addresses
incomplete and missing performance values in factorization computations, serving as a form
of tensor completion typically used in data imputation [21, 26, 27]). Inspired by the recent
advancements of generative models [28, 29], which are capable of generating data based on
patterns learned during training and have revolutionized simulation methodologies to be more
flexible and cost-effective, our study delves into exploring their potential of addressing the
data sparsity issue. We operate under the foundational assumption that, if learning patterns
can be identified within the multidimensional learner model, they can be effectively simulated
and generated using generative models, facilitating scalable data augmentation. Consequently,
current research was guided by following two Research Questions:
• RQ 1: What is the most effective method for integrating tensor factorization and genera-
tive models to develop a systematic framework that proficiently imputes and augments
sparse learning performance data?
• RQ 2: In the context of simulating learning performance data, how do Generative Adver-
sarial Network (GAN) and Generative Pre-trained Transformer (GPT) models compare in
terms of effectiveness and accuracy?
2. Methods
2.1. Dataset
Our study investigated a dataset derived from the AutoTutor ITS, focusing on learning per-
formance in reading comprehension. This dataset originates from lessons developed for the
Center for the Study of Adult Literacy (CSAL) [30, 31], specifically the ’Cause and Effect’ lesson,
involving 118 participants. The lesson design incorporates three levels of question difficulty:
medium (M), easy (E), and hard (H). There are 9 medium-difficulty questions, 10 easy questions,
and 10 hard questions. Notably, the distribution of learners across these difficulty levels varies
within the lesson. Upon completing the medium difficulty level, learners are either advanced to
the hard level or redirected to the easy level, depending on their performance, thus providing a
tailored learning pathway.
2.2. The Systematic Simulation Framework
We propose a systematic simulation framework, 3DG, illustrated in Figure 1. This framework
begins by structuring the initial learning performance data, sourcing from real-world learner-
ITS interactions, into a three-dimensional tensor by dimensions of learners, questions, and
attempts. As depicted in the sparse cube space in Figure 1, filled cubes represent recorded values
of learning performance, while transparent cubes indicate missing values. Tensor completion
Figure 1: The 3-dimensions, Densification, and Generation (3DG) systematic simulation framework.
(based on tensor factorization) is then utilized, converting the sparse tensor to a densified one.
The densified tensor provides invaluable information in identifying various learning patterns,
which aids in dividing the tensor into sub-tensors by categorizing distinct learning patterns.
Subsequently, generative models are harnessed to simulate additional data samples for enriching
the original dataset based on each specific learning pattern. The entire operation is encapsulated
for scalable simulation sampling and ultimately offers a comprehensive dataset incorporating
both imputed and augmented data. This framework was developed to address RQ 1. More
detailed methods within this framework are described in the following subsections.
2.3. Tensor Completion for Data Imputation
The three-dimensional tensor 𝒯 , representing the learning process, is defined as 𝒯 ∈
𝑅𝑈 ×𝑁 ×𝑀 , where the 𝑈 = 𝑚𝑎𝑥(1, 2, 3, · · · , 𝑢) is the maximum number of learners, 𝑁 =
𝑚𝑎𝑥(1, 2, 3, · · · , 𝑛) the maximum number of questions, and 𝑀 = 𝑚𝑎𝑥(1, 2, 3, · · · , 𝑚) the
maximum number of attempts. Each element 𝜏𝑢𝑖𝑗 of 𝒯 indicates the performance variable
of learner 𝑙𝑢 on question 𝑞𝑖 at the attempt 𝑎𝑗 . For instance, in the CSAL AutoTutor context,
a binary variable 𝜏 𝑢𝑖𝑗 = {0, 1} is used, where 1 signifies a correct answer and 0 denotes an
incorrect one. We model the tenor 𝒯 as a factorization of two lower dimensional components:
1) a learner latent matrix 𝒰 of size 𝑈 × 𝐾 (𝐾 represents the set of latent features in tensor
factorization), which captures learner-related latent features matrix/space (such as abilities and
learning-related features); and 2) a latent tensor 𝒱 of size 𝐾 × 𝑀 × 𝑁 , representing the learner
knowledge in terms of latent features during question attempts. The approximated tensor 𝒯ˆ is
obtained by the following formula:
𝒯ˆ ≈ 𝒰 × 𝒱 (1)
where 𝒰 can be interpreted as the latent feature space encapsulating learner-related effects,
reflecting characteristics such as individual abilities/features and learning preferences. On the
other hand, the tensor 𝒱 represents the interaction between attempts and question-related
(knowledge acquisition) effects, adapting to various learner features.
2.4. Scalable Simulation based on Generative Models for Data Augmentation
To answer RQ 2, we used two generative models, GAN (Generative Adversarial Network) and
GPT (Generative Pre-trained Transformer), to facilitate scalable simulations that are tailored to
individual learning patterns. According to [32, 33], GAN model is uniquely structured with a
dual-network architecture comprising a generator and a discriminator. This architecture enables
GAN model to excel in generating high-quality synthetic data. In comparison, GPT model is
distinguished by its use of transformer architecture, which empowers it to generate data that is
not only contextually relevant but also maintains a high degree of coherence [34, 35].
Before initiating the simulations, we employ a clustering algorithm (i.e., K-means++) to
categorize individual learning patterns based on similarities in learners’ performance. The
learners-attempts matrix slice extracted from the 𝒯ˆ , encapsulates the probability-based knowl-
edge states associated with the performance on the 𝑛th question 𝑞𝑛 , for all 𝑈 learners over 𝑀
attempts. In our analysis, we employ the “power law learning curve", a model widely recognized
in educational and training research [36, 37, 38], to fit the learning performance with increasing
attempts. In the power-law formula 𝑌 = 𝑎𝑋 𝑏 , the 𝑌 represents the learning performance,
quantified as the probability of producing correct answers, and 𝑋 is the number of opportunities
to practice a skill or attempt. The parameter 𝑎 indicates the measurement of the learner’s initial
ability or prior knowledge, and 𝑏 represents the learning rate at which the learner acquires
knowledge through practice. We employ K-means++ [39, 40] to cluster the distribution of two
model parameters (𝑎 and 𝑏), which assists in identifying distinct individual learning patterns.
As illustrated in Figure 2, the architecture of Generative Adversarial Network (GAN) consists
of two distinct neural networks: the Generator and the Discriminator. The Generator, often a
type of neural network like a convolutional neural network (CNN), is designed to create synthetic
data samples. It is denoted as 𝐺(·). The Discriminator, typically another neural network which
can also be a CNN (though its structure may vary based on the specific application), is tasked with
evaluating whether the data samples are real (authentic data) or fabricated by the Generator.
It is denoted as 𝐷(·). In the process (Figure 2), the Generator starts with a noise sample,
usually drawn from a Gaussian distribution, which has dimensions compatible with the original
data distribution. This noise sample serves as the initial input for the Generator, resulting
in 𝑆𝑖𝑚𝑢𝑙𝑎𝑡𝑒𝑑 𝐷𝑎𝑡𝑎. Both this 𝑆𝑖𝑚𝑢𝑙𝑎𝑡𝑒𝑑 𝐷𝑎𝑡𝑎 and the 𝑅𝑒𝑎𝑙 𝐷𝑎𝑡𝑎 are then fed into the
Discriminator 𝐷(·). The Discriminator’s role is to discern whether each sample is real or
simulated. Concurrently, the Generator 𝐺(·) is trained to progressively reduce the difference
between the distributions of the real and simulated data through iterative tuning.
Considering the limitations of using purely numerical values for interoperability and the
enhanced semantic understanding that detailed descriptions provide, we have developed a mixed-
based prompt approach for GPT-4 (illustrated in Figure 3). The prompt strategy integrates
original matrix data with interpretive text, thereby enriching the context and interpretability
of the data. Additionally, it incorporates the Chain-of-Thought (CoT) prompting technique
[41], which involves appending guiding phrases such as ’Let’s think step by step’ at the end
of the prompt to facilitate a more structured analytical process. Specifically, the constructed
prompt includes comprehensive elements such as the reading material being analyzed, detailed
information about the questions (including their answers), and the learners-attempts matrix
data, complete with descriptive information about both its format and entries. Subsequently, a
simulation request prompts GPT-4 to integrate the numerical and textual data in a coherent and
Figure 2: Diagram of using generative adversarial network (GAN) model for data simulation.
Figure 3: Diagram of using generative pre-trained transformer (GPT) model for data simulation.
insightful manner, ultimately driving the execution of a simulation. During the optimization
process, these prompts are iteratively refined and adjusted to efficiently yield results that align
with our specified objectives.
3. Results
As illustrated in Table 1, the original dataset exhibits sparsity levels ranging from 80% to 85%
(as determined by calculating the proportion of missing values to the total number of entries).
By iteratively tuning the latent feature range [1, 20] in tensor factorization algorithms, we
identified the optimal number of latent features (𝐾) as 6 for both Lesson (M) and Lesson (H),
and 4 for Lesson (E). The optimal 𝐾 value was derived by averaging results from multiple
trainings with optimized 𝐾 values in tensor factorization.
These findings suggest that tensor completion (based on tensor factorization) can efficiently
impute missing values in the original sparse performance data, notably for unexplored questions
and attempts. This enhancement is crucial for facilitating more comprehensive analysis and
modeling in Intelligent Tutoring Systems (ITSs). The latent features, closely associated with
learner-specific characteristics during the learning process, are captured with nuanced detail,
Table 1
Results about the sparsity levels and latent features by tensor completion. M, H, and E denote Medium,
Hard, and Easy lesson levels, respectively.
Dataset Sparsity Level (Original) 𝐾 (Latent Features)
Lesson (M) 84.02% 6
Lesson (H) 85.45% 6
Lesson (E) 81.25% 4
particularly in the context of reading comprehension. Further research is imperative to fully
understand the underlying physical essence of these latent features.
The distributions of parameters 𝑎 and 𝑏 are illustrated in Figure 4 and Figure 5, respectively.
These figures visualize the parameter distributions from a example cluster data set with an
original size of 20, and exhibit simulations in increments of 1000, with total sizes ranging from
1000 to 20000.
Figure 4 demonstrates the distributions of parameters 𝑎, which is used to represent the
learner’s initial ability or prior knowledge. Figure 4a shows the distribution of the parameter 𝑎
obtained by GAN simulation. As the sample size increases, the range of parameter 𝑎 from the
simulation sample mostly falls within the original range of parameter 𝑎, although it exhibits
a longer tail distribution extending beyond the original maximum value of parameter 𝑎. The
distribution of the parameter 𝑎 obtained by GPT-4 simulation is illustrated in Figure 4b. Unlike
those obtained from GAN simulation, the range of parameter 𝑎 values here extends beyond
the original range, which is particularly evident as the simulated sample size increases. This
(a) Distribution of parameter 𝑎 by GAN simulation.
(b) Distribution of parameter 𝑎 by GPT-4 simulation.
Figure 4: Distribution of parameter 𝑎 by simulation.
suggests that the initial learning ability in GPT-4 simulated samples exhibits more variability
and divergence from the original data compared to those from GAN simulation.
Then, Figure 5 demonstrates the distributions of parameters 𝑏 as derived from both GAN and
GPT-4 simulations. The parameter 𝑏 represents the learning rate, which reflects how quickly
a learner acquires knowledge through practice. The GAN simulation produces a narrower
range of parameter 𝑏 values, especially in terms of maximum and minimum values, when
compared to the original dataset, as depicted in Figure 5a. With increasing sample size, this
range generally maintains a consistent pattern. On the other hand, the GPT-4 simulation,
as shown in Figure 5b, demonstrates a broader range for parameter 𝑏, extending beyond the
original scope. This contrast suggests that GPT-4 simulation may capture a wider variability in
learning rates compared to GAN simulation.
(a) Distribution of parameter 𝑏 by GAN simulation.
(b) Distribution of parameter 𝑏 by GPT-4 simulation.
Figure 5: Distribution of parameter 𝑏 by simulation.
4. Discussion and Conclusion
This paper proposed the 3DG systematic simulation framework based on generative models
(particularly GAN and GPT) to address data sparsity challenges in learning performance data
within intelligent tutoring systems (ITS). The framework involves representing learner data from
problem-solving step attempts as a three-dimensional tensor with the axes of learners, questions,
and attempts. Tensor completion, based on tensor factorization, is then utilized to impute missing
performance data entries, generating a dense tensor. Such imputation computation leverages
the similarities in learner performance across various questions and attempts, capturing the
sequential and temporal dynamics of learning [42, 43]. We have demonstrated the integration
of generative models, including GAN and GPT-4 for creating scalable, individualized learning
simulations aimed at enhancing learner models for personalized instruction. Our comparative
analysis reveals that GAN surpasses GPT-4 in terms of reliability for scalable simulations.
Overall, the GAN simulations demonstrate a narrower and more consistent range of values
for parameters 𝑎 and 𝑏, indicating higher reliability for scalable simulations compared to
the broader value range exhibited by the GPT-4 simulations. The mechanism for the GPT-
4 simulations, refined through iterative optimization of GPT-4 prompts, involves selecting
random values from a flat array of the original data. These values are then adjusted to match
base probabilities, preserving the overall data distribution while facilitating the creation of
an expanded dataset. Although valuable in computational simulations, this method generally
underperforms in numerical computing compared to deep learning models, as demonstrated by
the GAN’s performance in this study.
Our findings shed light on the potential use of GPT-4 in simulating learner performance rep-
resented through numerical values in future research. Firstly, employing mixed-based prompts
improves interoperability with numerical data, thus enhancing the efficiency of subsequent
modeling and simulation computations. Secondly, the Chain-of-Thought (CoT) prompting tech-
nique delineates the steps for the simulation task, effectively directing GPT-4 in its reasoning
process. This includes a structured approach comprising: Understanding the Existing Matrix,
Distribution Analysis, Clustering Information, and Simulation Process. Thirdly, the
computational power of GPT-4 in modeling and simulation is attributed to its capabilities in
self-search, self-programming, and self-computing, all of which are facilitated by prompt engi-
neering. This significantly enhances its utility in data analysis and modeling for future research
endeavors. However, integrating GPT-4 with numerical computation presents fundamental
challenges, as we discuss in the following section.
5. Limitations and Future Works
The capability of GPT-4 in performing deep learning tasks involving numerical computations
remains insufficient, primarily due to the intrinsic limitations of large language models and
platform contraints. Future research could productively explore the integration of GAN with GPT
models, aiming to improve their interoperability and computational capabilities. Furthermore,
the degree of sparsity in the original performance data, particularly when formatted as a
tensor, significantly impacts the performance of generative models. Therefore, investigating the
sensitivity and robustness of tensor completion methods in response to different levels of data
sparsity presents an important avenue for future studies. Such investigations are crucial for
better integrating large language models within Intelligent Tutoring Systems (ITSs), potentially
leading to more refined and effective educational tools.
6. Acknowledgements
We extend our sincere gratitude to Prof. Philip I. Pavlik Jr. from the University of Memphis and
Prof. Shaghayegh Sahebi from the University at Albany - SUNY for their expert guidance on
tensor factorization method.
References
[1] J. R. Anderson, C. F. Boyle, B. J. Reiser, Intelligent tutoring systems, Science 228 (1985)
456–462.
[2] A. T. Corbett, K. R. Koedinger, J. R. Anderson, Intelligent tutoring systems, in: Handbook
of human-computer interaction, Elsevier, 1997, pp. 849–874.
[3] K. VanLehn, The relative effectiveness of human tutoring, intelligent tutoring systems,
and other tutoring systems, Educational psychologist 46 (2011) 197–221.
[4] A. C. Graesser, X. Hu, R. Sottilare, Intelligent tutoring systems, in: International handbook
of the learning sciences, Routledge, 2018, pp. 246–255.
[5] T.-N. Nguyen, Predicting student performance in an intelligent tutoring system, Ph.D.
thesis, Stiftung Universität Hildesheim, 2011.
[6] S. Pandey, G. Karypis, A self-attentive model for knowledge tracing, arXiv preprint
arXiv:1907.06837 (2019).
[7] N. Zhang, Y. Du, K. Deng, L. Li, J. Shen, G. Sun, Attention-based knowledge tracing with
heterogeneous information network embedding, in: Knowledge Science, Engineering
and Management: 13th International Conference, KSEM 2020, Hangzhou, China, August
28–30, 2020, Proceedings, Part I 13, Springer, 2020, pp. 95–103.
[8] C. Wang, S. Sahebi, S. Zhao, P. Brusilovsky, L. O. Moraes, Knowledge tracing for complex
problem solving: Granular rank-based tensor factorization, in: Proceedings of the 29th
ACM Conference on User Modeling, Adaptation and Personalization, 2021, pp. 179–188.
[9] X. Wang, S. Zhao, L. Guo, L. Zhu, C. Cui, L. Xu, Graphca: Learning from graph counter-
factual augmentation for knowledge tracing, IEEE/CAA Journal of Automatica Sinica 10
(2023) 2108–2123.
[10] P. I. Pavlik Jr, L. G. Eglington, L. Zhang, Automatic domain model creation and improve-
ment., Grantee Submission (2021).
[11] L. G. Eglington, P. I. Pavlik Jr, How to optimize student learning using student models
that adapt rapidly to individual differences, International Journal of Artificial Intelligence
in Education 33 (2023) 497–518.
[12] L. Zhang, P. I. Pavlik Jr, X. Hu, J. L. Cockroft, L. Wang, G. Shi, Exploring the individual
differences in multidimensional evolution of knowledge states of learners, in: International
Conference on Human-Computer Interaction, Springer, 2023, pp. 265–284.
[13] T. Wang, F. Ma, J. Gao, Deep hierarchical knowledge tracing, in: Proceedings of the 12th
international conference on educational data mining, 2019.
[14] W. Lee, J. Chun, Y. Lee, K. Park, S. Park, Contrastive learning for knowledge tracing, in:
Proceedings of the ACM Web Conference 2022, 2022, pp. 2330–2338.
[15] A. Kossiakoff, W. N. Sweet, S. J. Seymour, S. M. Biemer, Systems engineering principles
and practice, volume 83, John Wiley & Sons, 2011.
[16] M. Baudin, A. Dutfoy, B. Iooss, A.-L. Popelin, Openturns: An industrial software for
uncertainty quantification in simulation, in: Handbook of uncertainty quantification,
Springer, 2017, pp. 2001–2038.
[17] E. Acar, D. M. Dunlavy, T. G. Kolda, M. Mørup, Scalable tensor factorizations for incomplete
data, Chemometrics and Intelligent Laboratory Systems 106 (2011) 41–56.
[18] C. Shorten, T. M. Khoshgoftaar, A survey on image data augmentation for deep learning,
Journal of big data 6 (2019) 1–48.
[19] T. Emmanuel, T. Maupong, D. Mpoeleng, T. Semong, B. Mphago, O. Tabona, A survey on
missing data in machine learning, Journal of Big Data 8 (2021) 1–37.
[20] T. Liu, J. Fan, Y. Luo, N. Tang, G. Li, X. Du, Adaptive data augmentation for supervised
learning over missing data, Proceedings of the VLDB Endowment 14 (2021) 1202–1214.
[21] N. Thai-Nghe, L. Drumond, T. Horváth, A. Krohn-Grimberghe, A. Nanopoulos, L. Schmidt-
Thieme, Factorization techniques for predicting student performance, in: Educational
recommender systems and technologies: Practices and challenges, IGI Global, 2012, pp.
129–153.
[22] A. Krizhevsky, I. Sutskever, G. E. Hinton, Imagenet classification with deep convolutional
neural networks, Advances in neural information processing systems 25 (2012).
[23] S. Sahebi, Y.-R. Lin, P. Brusilovsky, Tensor factorization for student modeling and perfor-
mance prediction in unstructured domain., International Educational Data Mining Society
(2016).
[24] T.-N. Doan, S. Sahebi, Rank-based tensor factorization for student performance prediction,
in: 12th International Conference on Educational Data Mining (EDM), 2019.
[25] S. Zhao, C. Wang, S. Sahebi, Modeling knowledge acquisition from multiple learning
resource types, arXiv preprint arXiv:2006.13390 (2020).
[26] N. Thai-Nghe, T. Horváth, L. Schmidt-Thieme, et al., Factorization models for forecasting
student performance., in: EDM, Eindhoven, 2011, pp. 11–20.
[27] X. Chen, Z. He, Y. Chen, Y. Lu, J. Wang, Missing traffic data imputation and pattern
discovery with a bayesian augmented tensor factorization model, Transportation Research
Part C: Emerging Technologies 104 (2019) 66–77.
[28] M. Jovanovic, M. Campbell, Generative artificial intelligence: Trends and prospects,
Computer 55 (2022) 107–112.
[29] D. Baidoo-Anu, L. Owusu Ansah, Education in the era of generative artificial intelligence
(ai): Understanding the potential benefits of chatgpt in promoting teaching and learning,
Available at SSRN 4337484 (2023).
[30] A. C. Graesser, Z. Cai, W. O. Baer, A. M. Olney, X. Hu, M. Reed, D. Greenberg, Reading
comprehension lessons in autotutor for the center for the study of adult literacy, in:
Adaptive educational technologies for literacy instruction, Routledge, 2016, pp. 288–293.
[31] Y. Fang, K. Shubeck, A. Lippert, Q. Chen, G. Shi, S. Feng, J. Gatewood, S. Chen, Z. Cai,
P. Pavlik, et al., Clustering the learning patterns of adults with low literacy skills interacting
with an intelligent tutoring system., Grantee Submission (2018).
[32] I. Goodfellow, J. Pouget-Abadie, M. Mirza, B. Xu, D. Warde-Farley, S. Ozair, A. Courville,
Y. Bengio, Generative adversarial nets, Advances in neural information processing systems
27 (2014).
[33] I. Goodfellow, J. Pouget-Abadie, M. Mirza, B. Xu, D. Warde-Farley, S. Ozair, A. Courville,
Y. Bengio, Generative adversarial networks, Communications of the ACM 63 (2020)
139–144.
[34] A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, Ł. Kaiser, I. Polo-
sukhin, Attention is all you need, Advances in neural information processing systems 30
(2017).
[35] A. Radford, K. Narasimhan, T. Salimans, I. Sutskever, et al., Improving language under-
standing by generative pre-training (2018).
[36] A. Newell, P. S. Rosenbloom, Mechanisms of skill acquisition and the law of practice.,
Technical Report, CARNEGIE-MELLON UNIV PITTSBURGH PA DEPT OF COMPUTER
SCIENCE, 1980.
[37] H. Cen, K. Koedinger, B. Junker, Learning factors analysis–a general method for cognitive
model evaluation and improvement, in: International conference on intelligent tutoring
systems, Springer, 2006, pp. 164–175.
[38] R. DeKeyser, Skill acquisition theory, in: Theories in second language acquisition,
Routledge, 2020, pp. 83–104.
[39] D. Arthur, S. Vassilvitskii, K-means++ the advantages of careful seeding, in: Proceedings of
the eighteenth annual ACM-SIAM symposium on Discrete algorithms, 2007, pp. 1027–1035.
[40] B. Bahmani, B. Moseley, A. Vattani, R. Kumar, S. Vassilvitskii, Scalable k-means++, arXiv
preprint arXiv:1203.6402 (2012).
[41] J. Wei, X. Wang, D. Schuurmans, M. Bosma, F. Xia, E. Chi, Q. V. Le, D. Zhou, et al., Chain-
of-thought prompting elicits reasoning in large language models, Advances in Neural
Information Processing Systems 35 (2022) 24824–24837.
[42] C. M. Conway, M. H. Christiansen, Sequential learning in non-human primates, Trends in
cognitive sciences 5 (2001) 539–546.
[43] C. M. Conway, Sequential Learning, Springer US, Boston, MA, 2012, pp. 3047–3050.