Pith. sign in

REVIEW 4 major objections 5 minor 54 references

DASKT: A Dynamic Affect Simulation Method for Knowledge Tracing

T0 review · 4 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Students' affective states can be simulated from ordinary answer logs, and injecting them into knowledge tracing improves prediction of future performance.

desk verdict A novel affect-aware KT architecture undone, as written, by future-looking affect features; the core performance claim is not credible yet. read the letter →

arxiv 2502.10396 v1 pith:5OHNYYMK submitted 2025-01-18 cs.CY cs.AIcs.LG

Xinjie Sun , Kai Zhang , Qi Liu , Shuanghong Shen , Fei Wang , Yuxiang Guo , Enhong Chen This is my paper · ORCID
classification cs.CYcs.AIcs.LG
keywords knowledgetracingaffectivecomputingaffectsimulationstudentperformancepredictiongraphattentionnetworkLSTMconfidence-interest-efforteducationaldatamining
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

Knowledge tracing predicts a student's next answer from their exercise history, and the paper claims that knowing how the student feels—frustrated, concentrated, bored, or confused—makes that prediction better and more interpretable. DASKT is a pipeline that derives three affective factors (confidence, interest, effort) from ordinary behavioral logs such as answer correctness, response time, participation frequency, attempts, and inter-question intervals; clusters students into four affective states; and models how affect changes over time using a graph attention network. On two public datasets, ASSIST2012 and ASSISTchall, the reported area under the ROC curve (AUC) is 0.7925 and 0.8529 respectively, above all twelve baselines, and the simulated affect agrees with an external affect detector on 72.74% to 80.29% of questions on ASSIST2012. If the claim holds, affect can be brought into intelligent tutoring systems at no extra sensing cost.

What carries the argument

The load-bearing mechanism is the affect trajectory graph built on interval-computed affect labels. DASKT first normalizes each student's confidence, interest, and effort vectors (the MAF module), clusters them into four affect states with K-means, and assigns an affect label to each interval of LSeg exercises (the ICA module). These interval labels become nodes of a graph in which each node is connected to its predecessor and successor, and a two-layer graph attention network—AT-GAT followed by A-GAT—produces a smoothed, coherent affect representation at each time step. That representation is concatenated with the problem, knowledge-concept, and response embeddings and fed into a long short-term memory (LSTM) network, so the student's knowledge state is updated under the influence of the simulated affect.

What would settle it

Recompute every affective factor (confidence, interest, effort) using only interactions strictly before the predicted time, retrain DASKT on the same data splits, and compare AUC; if the advantage disappears, the reported gains come from information leakage rather than from modeling affect.

Watch

Extended reading notes

Core claim

The paper's central claim is that affect in online learning can be simulated from non-affect-oriented behavioral logs, without sensors, surveys, or gold-standard affect labels, and that the simulated affect improves knowledge tracing in two ways. On prediction, DASKT reports AUC of 0.7925 on ASSIST2012 and 0.8529 on ASSISTchall, beating all twelve baselines; on interpretability, the affect-conditioned knowledge states show the expected pattern that concentrated students gain more and frustrated or bored students lose more for the same answers. The simulated affect also agrees with the affect detector on ASSIST2012 for 72.74% to 80.29% of questions, and feeding the simulated affect into the DKT and GKT models improves their AUC, which the paper reads as evidence that the affect signal is accurate and transferable.

Load-bearing premise

The load-bearing premise is that the affect features used to predict an answer are honest predictors, even though they are computed from statistics over the student's entire exercise history, including the answer being predicted; recomputing them from only past interactions might erase the reported gains.

Editorial extensions

If this is right

  • Affect simulation can run at scale from logs alone, so intelligent tutoring systems without affect labels or sensors can still incorporate affect into their student models.
  • Knowledge-state estimates become affect-sensitive: the paper shows positive affect raises mastery trajectories and negative affect lowers them, which makes predicted mastery more interpretable and opens the door to affect-targeted interventions.
  • The affect component transfers across architectures: feeding the simulated affect into DKT and GKT raises their AUC on ASSIST2012, indicating it is not tied to a single model design.
  • Long-sequence prediction benefits most, since the largest reported gain is on ASSISTchall, a longitudinal dataset, suggesting affect matters more when learning histories are long.
  • Without the affect modules the model's prediction quality drops (AUC from 0.7925 to 0.7404 on ASSIST2012 in the full ablation), so the affect pathway is load-bearing for the reported performance.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Editorial inference: The reported gains may shrink if affective factors are recomputed causally, since the current formulas use whole-sequence averages and therefore include information from the outcome being predicted; a strictly causal version would separate true affect signal from leakage.
  • Editorial inference: The same confidence–interest–effort pipeline could be tested in other education tasks without affect labels, such as cognitive diagnosis or dropout prediction, though the paper only evaluates knowledge tracing.
  • Editorial inference: Segment length (LSeg) and the number of affect clusters are free parameters, so a sensitivity analysis would show how much of the gain is robust and how much depends on these choices.
  • Editorial inference: Because the affect cluster centers are learned from the training population, deploying DASKT on a new platform or school would likely require recalibration of the affect states to that population.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. The paper proposes DASKT, a knowledge tracing model that simulates four affective states (frustration, concentration, boredom, confusion) from behavioral records without explicit affect labels. Affective factors (confidence, interest, effort) are computed from whole-sequence statistics, clustered into four states via K-means, and then processed by a two-layer graph attention network to form a dynamic affect trajectory. The affect embedding is concatenated with problem, KC, and response embeddings into an LSTM to predict future performance. Experiments on ASSIST2012 and ASSISTchall report AUC of 0.7925 and 0.8529, outperforming 12 baselines, with an ablation study attributing roughly five AUC points to the affect components.

Significance. If the claims held, the paper would offer a low-cost, label-free way to incorporate affect into knowledge tracing and improve both accuracy and interpretability. The manuscript also states that data and code are released, which is a strength for reproducibility. However, the central performance claim is undermined by a severe information-leakage problem in the affective feature construction, and the affect validation is incomplete and partly ambiguous. As written, the reported gains cannot be attributed to a causal effect of affect, so the significance of the contribution is not established.

major comments (4)
  1. [Section 3.2, Eqs. (1)-(9); Section 4.2, Eq. (12); Section 4.3, Eq. (16)] The affective factors are computed from whole-sequence statistics that include future responses. Specifically, Eq. (1) defines A_{s_i}^k as the average correctness over all occurrences of KC k in the student's entire sequence; Eq. (4) uses |el_{s_i}|, the full sequence length, to define participation rate; Eqs. (7)-(8) sum attempts and intervals over the whole sequence. These quantities have no time index and are therefore identical at every time step, encoding information about responses at time t and later. Eq. (12) then assigns an affect state to each segment using these same global factors, and Eq. (16) concatenates the resulting affect embedding into the LSTM input at time t to predict r_{t+1}. Consequently, the model has a direct path to future information, so the performance gains in Table 4 and the ablation findings in Table 5 may be leakage artifacts rather than evidence of a causal affect effect. The authors must reformulate the affective factors so that the value at time t depends only on interactions before t, or they must demonstrate that the reported results are unchanged under such a causal formulation.
  2. [Section 4.2, Eq. (11)] The paper never specifies how the four K-means clusters are mapped to the emotion labels frustration, concentration, boredom, and confusion. Section 3.3.2 simply states that the affect embedding has four states, and Section 5.4 compares DASE to the affect detector of Pardos et al. [35]. Without an explicit mapping procedure (for example, matching clusters to detector labels on the training set), the consistency rates in Table 3 and the qualitative affect interpretations in Figures 4 and 5 are not reproducible, and the claim that no affect labels are used in training is difficult to verify.
  3. [Table 3] The per-affect AUC values (e.g., 15.27, 16.22, 20.17, 21.37) are implausible for an AUC metric, which must lie between 0 and 1 (or 0 and 100). The relationship between these per-affect values and the "Total" row (73.16, 75.24, etc.) is unexplained, and the notation "↑0.12" is ambiguous without a baseline definition. Moreover, the validation compares DASE with the affect detector on the same dataset used to develop the model; if the cluster-to-label mapping is chosen to maximize agreement with the detector, the consistency rates would be inflated. Please clarify the metric definitions, the mapping procedure, and whether detector labels are used only for evaluation.
  4. [Tables 4 and 5] No standard deviations or confidence intervals are reported despite the stated use of 5-fold cross-validation. The significance asterisks in Table 4 are not accompanied by variance information, and the ablation in Table 5 also lacks error bars. Given the leakage issue in the affective features, the reported improvements cannot be judged as statistically reliable, and the comparisons with baselines may not be meaningful as presented.
minor comments (5)
  1. [Section 3.2.1, Eq. (1)] The notation AT_{s_i}^k = att_{s_i}^{t_k} is unclear because t_k is not defined; it appears to mix a time index with a KC index, and the equation should specify how response time is aggregated.
  2. [Section 4.2, Eq. (12)] The phrase "|sn| is the set of all students in the training and test sets" is confusing because Eq. (11) uses |sm| for the training set; the text should clarify how test students are assigned to clusters and whether any test-set statistics are used in building the clusters.
  3. [Section 5.1 and Section 5.4] The statement "no real affect labels were used in training" should be reconciled with the use of the affect detector in Section 5.4; please state explicitly whether the detector labels are used only for evaluation or also for choosing the cluster-to-emotion mapping.
  4. [Section 2.1 heading] The heading "Konwledge Tracing" contains a typo; it should read "Knowledge Tracing."
  5. [Figure 2] Figure 2 is dense and the subfigure labels (a), (b), (c) are not clearly separated in the rendered PDF, making it hard to map the text description to the diagram.

Circularity Check

1 steps flagged · score 6.0 of 10

Whole-sequence affective factors fed into the LSTM at time t contain the response label r_{t+1}, so the reported prediction gains may be a leakage artifact rather than an affect effect.

  1. fitted input called prediction [Sec. 3.2.1 Eqs. (1)-(3); Sec. 4.1 Eq. (10); Sec. 4.2 Eq. (12); Sec. 4.3 Eq. (16); Sec. 4.4 Eq. (17)]
    "Asi k = 1/N si_k \sum_{n=1}^{C si_k} {rsi_nk == 1}, ... rsi_nk is the label indicating whether the student si answered correctly on the k-th KC in their own exercise sequence ... Conf si_k = concat( \sum (Asi_k - Ak), \sum (AT si_k - ATk)), ... ht = LST M(pt ⊕ kct ⊕ A-GAT(2)_n ⊕ rt), ... yt+1 = σ(wtyt + bt)"

    Equation (1) averages the response labels r over all occurrences of a KC in the student's 'exercise sequence' with no time cutoff, and Eq. (3) places that whole-sequence average into the confidence factor. Via Eqs. (10), (12), (13), and (15), this factor becomes the affect representation A-GAT that Eq. (16) concatenates into the LSTM input at time t, from which Eq. (17) predicts r_{t+1}. Therefore the target label enters the predictor through the affect input whenever the next problem's KC occurs anywhere in the sequence, including at t+1 or later. The AUC gains in Table 4 and the ablation deltas in Table 5 may thus reflect leakage rather than a causal effect of affect.

full rationale

The only load-bearing circularity I can exhibit from the paper's own equations is the construction of the affect input. Eq. (1) defines accuracy A^{s_i}_k as the mean of r over all occurrences of KC k in the student's exercise sequence, with no restriction to t' < t; Eq. (3) packs this into confidence; Eq. (10) normalizes it into the affective factor vector; Eq. (12) assigns interval affect from these factors; Eqs. (13)-(16) carry that affect into the LSTM state h_t; and Eq. (17) predicts r_{t+1} from h_t. Thus the target label is part of the input by construction whenever the same KC appears at or after t, including the next problem. The reported improvements are therefore not clean evidence for a causal affect effect. I did not find a load-bearing self-citation chain: citations to the authors' own KT papers are for baselines and related work, and the affect-state choice is attributed to external work by Pardos et al. The RQ1 consistency check against the affect detector is not demonstrably circular because the paper does not describe how cluster indices are mapped to the four affect names, but it is also not independent evidence; I am not counting it as a separate scored step because no equation or described procedure shows the mapping is fitted to the detector. Score 6 reflects that the central performance claim partially reduces to a fitted whole-sequence statistic that contains the prediction target; if the authors recompute the affective factors strictly from past interactions, the circularity would be resolved.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The central performance claim rests on the ad hoc construction of affective factors and the post hoc labeling of clusters. The paper introduces no new physical or theoretical entities; the key burden is on the behavioral assumptions and the unknown LSeg and label assignment.

free parameters (4)
  • LSeg (interval segment length) = not reported
    Controls the granularity of affect assignment; example shows 5, but no value or sensitivity analysis is given in the experimental setup (Section 5.2).
  • Global behavioral averages (A_k, AT_k, PR, Att_k, IT) = computed from the training set
    These define confidence, interest, and effort as deviations from cohort means (Eq. 1-9); they are estimated from the data and, as written, over the full student sequence.
  • K-means cluster centers for affect states = 4 centers derived from training set
    Cluster centers in Eq. 11 are fit to training data; the mapping from cluster index to the named affects is not derived from data or theory.
  • Sequence length thresholds (100 for ASSIST2012, 500 for ASSISTchall) = 100/500
    Chosen threshold for segmenting long sequences into virtual students; no justification for the values.
assumptions (5)
  • domain assumption The four affect states used in the paper (frustration, concentration, boredom, confusion) are sufficient and appropriate for modeling affect in KT.
    Section 3.1 inherits the affect set from Pardos et al. [35] without independently validating the categorical scheme.
  • domain assumption Confidence, interest, and effort as defined in Section 3.2 capture the relevant affective factors.
    The formulas are based on Classical Test Theory and prior work [40,41], but no psychometric validation is provided.
  • domain assumption K-means with k=4 produces a meaningful partition of students into the four affect states.
    Section 4.2 uses K-means; the cluster labels are assigned after the fact and not part of the algorithm.
  • domain assumption Temporal coherence of affect can be captured by a graph attention network over adjacent time steps.
    Section 4.3, Eq. 14-15; relies on psychological studies [51-53] that affect changes are coherent.
  • ad hoc to paper The global statistics used in affect factor computation (Eq. 1-9) are legitimate for use in a time-series prediction setting.
    As written, these statistics use the full student sequence, including future responses, which would leak information into the KT prediction task.

how reviews work

0 comments
Cite this review

Pith. "Pith review of DASKT: A Dynamic Affect Simulation Method for Knowledge Tracing." pith.science (2026). https://pith.science/paper/5OHNYYMK

@misc{pith2026250210396,
  author       = {Pith},
  title        = {Pith review of: DASKT: A Dynamic Affect Simulation Method for Knowledge Tracing},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5OHNYYMK}},
  note         = {Machine review of arXiv:2502.10396}
}
read the original abstract

Knowledge Tracing (KT) predicts future performance by modeling students' historical interactions, and understanding students' affective states can enhance the effectiveness of KT, thereby improving the quality of education. Although traditional KT values students' cognition and learning behaviors, efficient evaluation of students' affective states and their application in KT still require further exploration due to the non-affect-oriented nature of the data and budget constraints. To address this issue, we propose a computation-driven approach, Dynamic Affect Simulation Knowledge Tracing (DASKT), to explore the impact of various student affective states (such as frustration, concentration, boredom, and confusion) on their knowledge states. In this model, we first extract affective factors from students' non-affect-oriented behavioral data, then use clustering and spatiotemporal sequence modeling to accurately simulate students' dynamic affect changes when dealing with different problems. Subsequently, {\color{blue}we incorporate affect with time-series analysis to improve the model's ability to infer knowledge states over time and space.} Extensive experimental results on two public real-world educational datasets show that DASKT can achieve more reasonable knowledge states under the effect of students' affective states. Moreover, DASKT outperforms the most advanced KT methods in predicting student performance. Our research highlights a promising avenue for future KT studies, focusing on achieving high interpretability and accuracy.

Figures

Figures reproduced from arXiv: 2502.10396 by the authors.

Figure 1
Figure 1. A toy example illustrating the mastery level of three knowledge concepts in the same exercise sequence for two students who have identical cognitive abilities at the ini￾tial stage. Student s1 experiences affect of frustration and boredom throughout the learning process, while Student s2 consistently maintains a state of concentration. researchers have recently incorporated educational theory into their approaches, … view at source ↗
Figure 2
Figure 2. The main structure of our DASKT model is presented. The diagram illustrates the method of obtaining affective factors and the process of calculating affect at intervals. Based on this, at time step t, the inputs include affect affectt, problem embedding Pt, knowledge concept of the problem KCt, student’s response rt, and knowledge states ht. In addition, we can also predict the student’s performance at time step t +… view at source ↗
Figure 3
Figure 3. An illustrative example of interval-based affect calculation is shown where a student’s sequence of 14 attempts is split into three interval segments, correlating to time intervals for answering 5 problems each. If the student ceases interaction, remaining segments are zero￾filled. Attempt counts vary according to the number of problems addressed. begins, based on the evaluation of the student’s previous performance… view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: This is a comparative diagram that illustrates the differences in a student’s knowledge states when com￾pleting a sequence of nine exercises, with and without affect involvement. The knowledge states is updated every three exercises, with numerals representing the valu…
Figure 5
Figure 5. Figure 5: This is an example demonstrated on the ASSISTchall dataset, revealing the impact of dynamic affective states on the changes in knowledge states and the performance of the DASKT model. We interpret future predictions by comparing the dynamic affective states in the lear…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

54 extracted references · 48 canonical work pages

  1. [35]

    Affective states and state tests: Investigating how affect throughout the school year predicts end of year learning outcomes,

    Z. A. Pardos, R. S. Baker, M. O. San Pedro, S. M. Gowda, and S. M. Gowda, “Affective states and state tests: Investigating how affect throughout the school year predicts end of year learning outcomes,” in Proceedings of the third international conference on learning analytics and knowledge, 2013, pp. 117–124

    work page 2013

  2. [2]

    Learning process-consistent knowledge tracing,

    S. Shen, Q. Liu, E. Chen, Z. Huang, W. Huang, Y. Yin, Y. Su, and S. Wang, “Learning process-consistent knowledge tracing,” in Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery & Data Mining , ser. KDD ’21. New York, NY, USA: Association for Computing Machinery, 2021, p. 1452–1460

    work page 2021

  3. [3]

    A survey of knowledge tracing,

    Q. Liu, S. Shen, Z. Huang, E. Chen, and Y. Zheng, “A survey of knowledge tracing,” arXiv preprint arXiv:2105.15106, 2021

    arXiv 2021

  4. [4]

    Deep knowledge tracing,

    C. Piech, J. Bassen, J. Huang, S. Ganguli, M. Sahami, L. J. Guibas, and J. Sohl-Dickstein, “Deep knowledge tracing,” Advances in neural information processing systems, vol. 28, 2015

    2015

  5. [5]

    Deep knowledge tracing and dynamic student classification for knowl- edge tracing,

    S. Minn, Y. Yu, M. C. Desmarais, F. Zhu, and J.-J. Vie, “Deep knowledge tracing and dynamic student classification for knowl- edge tracing,” in 2018 IEEE International conference on data mining (ICDM). IEEE, 2018, pp. 1182–1187

    work page 2018

  6. [6]

    Improving interpretability of deep sequential knowledge tracing models with question-centric cognitive representations,

    J. Chen, Z. Liu, S. Huang, Q. Liu, and W. Luo, “Improving interpretability of deep sequential knowledge tracing models with question-centric cognitive representations,” arXiv preprint arXiv:2302.06885, 2023

    arXiv 2023

  7. [8]

    Context-aware attentive knowledge tracing,

    A. Ghosh, N. Heffernan, and A. S. Lan, “Context-aware attentive knowledge tracing,” in Proceedings of the 26th ACM SIGKDD inter- national conference on knowledge discovery & data mining , 2020, pp. 2330–2339

    work page 2020

  8. [9]

    Learning consistent representations with temporal and causal enhancement for knowledge tracing,

    C. Huang, H. Wei, Q. Huang, F. Jiang, Z. Han, and X. Huang, “Learning consistent representations with temporal and causal enhancement for knowledge tracing,” Expert Systems with Appli- cations, p. 123128, 2024. JOURNAL OF LATEX CLASS FILES 13

    work page 2024

Show all 54 references
  1. [10]

    Graph-based knowl- edge tracing: modeling student proficiency using graph neural network,

    H. Nakagawa, Y. Iwasawa, and Y. Matsuo, “Graph-based knowl- edge tracing: modeling student proficiency using graph neural network,” in IEEE/WIC/ACM International Conference on Web In- telligence, 2019, pp. 156–163

    2019

  2. [11]

    Jkt: A joint graph convolutional network based deep knowledge tracing,

    X. Song, J. Li, Y. Tang, T. Zhao, Y. Chen, and Z. Guan, “Jkt: A joint graph convolutional network based deep knowledge tracing,” Information Sciences, vol. 580, pp. 510–523, 2021

    2021

  3. [12]

    Dynamic cognitive diagnosis: An educational priors-enhanced deep knowledge tracing perspective,

    F. Wang, Z. Huang, Q. Liu, E. Chen, Y. Yin, J. Ma, and S. Wang, “Dynamic cognitive diagnosis: An educational priors-enhanced deep knowledge tracing perspective,” IEEE Transactions on Learn- ing Technologies, vol. 16, no. 3, pp. 306–323, 2023

    2023

  4. [13]

    Monitoring student progress for learning process- consistent knowledge tracing,

    S. Shen, E. Chen, Q. Liu, Z. Huang, W. Huang, Y. Yin, Y. Su, and S. Wang, “Monitoring student progress for learning process- consistent knowledge tracing,” IEEE Transactions on Knowledge and Data Engineering, 2022

    2022

  5. [15]

    Rkt: relation-aware self-attention for knowledge tracing,

    S. Pandey and J. Srivastava, “Rkt: relation-aware self-attention for knowledge tracing,” in Proceedings of the 29th ACM International Conference on Information & Knowledge Management, 2020, pp. 1205– 1214

    2020

  6. [16]

    Affect and inference in bayesian knowledge tracing with a robot tutor,

    S. Spaulding and C. Breazeal, “Affect and inference in bayesian knowledge tracing with a robot tutor,” in Proceedings of the Tenth Annual ACM/IEEE International Conference on Human-Robot Interac- tion Extended Abstracts, 2015, pp. 219–220

    2015

  7. [17]

    Gikt: a graph-based interaction model for knowledge tracing,

    Y. Yang, J. Shen, Y. Qu, Y. Liu, K. Wang, Y. Zhu, W. Zhang, and Y. Yu, “Gikt: a graph-based interaction model for knowledge tracing,” in Machine Learning and Knowledge Discovery in Databases: European Conference, ECML PKDD 2020, Ghent, Belgium, September 14–18, 2020, Proceedin...

    2020

  8. [18]

    A self-attentive model for knowledge tracing,

    S. Pandey and G. Karypis, “A self-attentive model for knowledge tracing,” arXiv preprint arXiv:1907.06837, 2019

    1907 arXiv

  9. [20]

    Dynamic key-value memory networks with rich features for knowledge tracing,

    X. Sun, X. Zhao, B. Li, Y. Ma, R. Sutcliffe, and J. Feng, “Dynamic key-value memory networks with rich features for knowledge tracing,” IEEE transactions on cybernetics , vol. 52, no. 8, pp. 8239– 8245, 2021

    2021

  10. [21]

    Knowledge tracing with se- quential key-value memory networks,

    G. Abdelrahman and Q. Wang, “Knowledge tracing with se- quential key-value memory networks,” in Proceedings of the 42nd international ACM SIGIR conference on research and development in information retrieval, 2019, pp. 175–184

    2019

  11. [22]

    Heterogeneous features integration in deep knowledge tracing,

    L. P . Cheung and H. Yang, “Heterogeneous features integration in deep knowledge tracing,” pp. 653–662, 2017

    2017

  12. [23]

    qdkt: Question-centric deep knowledge tracing,

    S. Sonkar, A. E. Waters, A. S. Lan, P . J. Grimaldi, and R. G. Baraniuk, “qdkt: Question-centric deep knowledge tracing,” arXiv preprint arXiv:2005.12442, 2020

    2005 arXiv

  13. [25]

    Enhanc- ing knowledge tracing via adversarial training,

    X. Guo, Z. Huang, J. Gao, M. Shang, M. Shu, and J. Sun, “Enhanc- ing knowledge tracing via adversarial training,” in Proceedings of the 29th ACM International Conference on Multimedia, 2021, pp. 367– 375

    2021

  14. [26]

    Boosting neural cognitive diagnosis with student’s affective state modeling,

    S. Wang, Z. Zeng, X. Yang, K. Xu, and X. Zhang, “Boosting neural cognitive diagnosis with student’s affective state modeling,” in Proceedings of the AAAI Conference on Artificial Intelligence , vol. 38, no. 1, 2024, pp. 620–627

    2024

  15. [27]

    Measuring emotions in students’ learning and performance: The achievement emotions questionnaire (aeq),

    R. Pekrun, T. Goetz, A. C. Frenzel, P . Barchfeld, and R. P . Perry, “Measuring emotions in students’ learning and performance: The achievement emotions questionnaire (aeq),” Contemporary educa- tional psychology, vol. 36, no. 1, pp. 36–48, 2011

    2011

  16. [28]

    The influence of positive and negative emotions on physiological responses and memory task scores,

    M. T. Riviello, V . Capuano, G. Ombrato, I. Baldassarre, G. Cor- dasco, and A. Esposito, “The influence of positive and negative emotions on physiological responses and memory task scores,” in Recent Advances of Neural Network Models and Applications: Pro- ceedings of the 23rd...

    2014

  17. [29]

    Anyone else have this experience: Sharing the emotional labor of tracking data about me,

    M. Knittel, F. Kollig, A. Mason, and R. Wash, “Anyone else have this experience: Sharing the emotional labor of tracking data about me,” Proceedings of the ACM on Human-Computer Interaction, vol. 5, no. CSCW1, pp. 1–30, 2021

    2021

  18. [30]

    Deep self- attention network for facial emotion recognition,

    A. Gupta, S. Arunachalam, and R. Balakrishnan, “Deep self- attention network for facial emotion recognition,” Procedia Com- puter Science, vol. 171, pp. 1527–1534, 2020

    2020

  19. [31]

    A novel driver emotion recognition system based on deep ensemble classification,

    K. Zaman, S. Zhaoyun, B. Shah, T. Hussain, S. M. Shah, F. Ali, and U. S. Khan, “A novel driver emotion recognition system based on deep ensemble classification,” Complex & Intelligent Systems , pp. 1–26, 2023

    2023

  20. [32]

    Speech emotion recognition: a comprehensive survey,

    M. J. Al-Dujaili and A. Ebrahimi-Moghadam, “Speech emotion recognition: a comprehensive survey,” Wireless Personal Communi- cations, vol. 129, no. 4, pp. 2525–2561, 2023

    2023

  21. [33]

    Sentiment analysis and opinion mining on educational data: A survey,

    T. Shaik, X. Tao, C. Dann, H. Xie, Y. Li, and L. Galligan, “Sentiment analysis and opinion mining on educational data: A survey,” Natural Language Processing Journal, vol. 2, p. 100003, 2023

    2023

  22. [34]

    Blr: A multi-modal sentiment analysis model,

    Y. Yang, Y. Zhonglin, Z. Haixing, L. Gege, and C. Shujuan, “Blr: A multi-modal sentiment analysis model,” in International Conference on Artificial Neural Networks. Springer, 2023, pp. 466–478

    2023

  23. [36]

    Unfold- ing sentimental and behavioral tendencies of learners’ concerned topics from course reviews in a mooc,

    S. Liu, X. Peng, H. N. Cheng, Z. Liu, J. Sun, and C. Yang, “Unfold- ing sentimental and behavioral tendencies of learners’ concerned topics from course reviews in a mooc,” Journal of Educational Computing Research, vol. 57, no. 3, pp. 670–696, 2019

    2019

  24. [37]

    Tempo- ral emotion-aspect modeling for discovering what students are concerned about in online course forums,

    Z. Liu, C. Yang, S. R ¨udian, S. Liu, L. Zhao, and T. Wang, “Tempo- ral emotion-aspect modeling for discovering what students are concerned about in online course forums,” Interactive Learning Environments, vol. 27, no. 5-6, pp. 598–627, 2019

    2019

  25. [38]

    Towards an understanding of affect and knowledge from student interaction with an intelligent tutoring system,

    M. O. Z. San Pedro, R. S. d. Baker, S. M. Gowda, and N. T. Heffernan, “Towards an understanding of affect and knowledge from student interaction with an intelligent tutoring system,” in Artificial Intelligence in Education: 16th International Conference, AIED 2013, Memphis, TN...

    2013

  26. [39]

    The smell, the emotion, and the lebowski shock: What virtual education abroad cannot do?

    W. Liu, D. Sulz, and G. Palmer, “The smell, the emotion, and the lebowski shock: What virtual education abroad cannot do?.” Journal of Comparative and International Higher Education , vol. 14, no. 2, pp. 112–125, 2022

    2022

  27. [40]

    Diagnosing and acting on student affect: the tutor’s perspective,

    K. Porayska-Pomsta, M. Mavrikis, and H. Pain, “Diagnosing and acting on student affect: the tutor’s perspective,” User Modeling and User-Adapted Interaction, vol. 18, pp. 125–173, 2008

    2008

  28. [41]

    Knowledge elicitation methods for affect modelling in education,

    K. Porayska-Pomsta, M. Mavrikis, S. D’Mello, C. Conati, and R. S. Baker, “Knowledge elicitation methods for affect modelling in education,” International Journal of Artificial Intelligence in Education, vol. 22, no. 3, pp. 107–140, 2013

    2013

  29. [42]

    Classical test theory,

    R. F. DeVellis, “Classical test theory,” Medical care , pp. S50–S59, 2006

    2006

  30. [43]

    The relationship between response time and diagnostic accuracy,

    J. Sherbino, K. L. Dore, T. J. Wood, M. E. Young, W. Gaissmaier, S. Kreuger, and G. R. Norman, “The relationship between response time and diagnostic accuracy,” Academic Medicine , vol. 87, no. 6, pp. 785–791, 2012

    2012

  31. [44]

    Situational interest and physical activity in fitness testing: A need for pedagogical engineering,

    X. Zhu, “Situational interest and physical activity in fitness testing: A need for pedagogical engineering,” International Journal of Sport and Exercise Psychology, vol. 12, no. 1, pp. 76–89, 2014

    2014

  32. [45]

    Effort-based tutoring: An empirical approach to intelligent tutoring,

    I. Arroyo, H. Mehranian, and B. P . Woolf, “Effort-based tutoring: An empirical approach to intelligent tutoring,” in Educational data mining 2010. ERIC, 2010

    2010

  33. [46]

    Researching and practicing positive psychology in second/foreign language learn- ing and teaching: the past, current status and future directions,

    Y. Wang, A. Derakhshan, and L. J. Zhang, “Researching and practicing positive psychology in second/foreign language learn- ing and teaching: the past, current status and future directions,” Frontiers in Psychology, vol. 12, p. 731721, 2021

    2021

  34. [47]

    Positive emotions, self-regulatory capacity, and efl performance in chinese senior high school stu- dents,

    S. Li, H. Wu, and Y. Wang, “Positive emotions, self-regulatory capacity, and efl performance in chinese senior high school stu- dents,” Acta Psychologica, vol. 243, p. 104143, 2024

    2024

  35. [48]

    Negative affect improves the quality of memories: trading capacity for precision in sensory and working memory

    P . Spachtholz, C. Kuhbandner, and R. Pekrun, “Negative affect improves the quality of memories: trading capacity for precision in sensory and working memory.” Journal of Experimental Psychology: General, vol. 143, no. 4, p. 1450, 2014

    2014

  36. [49]

    Negative emotion reduces visual working memory recall variability: A meta-analytical review

    W. Xie, C. Ye, and W. Zhang, “Negative emotion reduces visual working memory recall variability: A meta-analytical review.” Emotion, vol. 23, no. 3, p. 859, 2023

    2023

  37. [50]

    K-means clustering-based kernel canonical correlation analysis for multimodal emotion recognition in human–robot interaction,

    L. Chen, K. Wang, M. Li, M. Wu, W. Pedrycz, and K. Hirota, “K-means clustering-based kernel canonical correlation analysis for multimodal emotion recognition in human–robot interaction,” JOURNAL OF LATEX CLASS FILES 14 IEEE Transactions on Industrial Electronics , vol. 70, no....

    2022

  38. [51]

    Emotional development

    L. J. Yarrow, “Emotional development.” American Psychologist , vol. 34, no. 10, p. 951, 1979

    1979

  39. [52]

    Emotion regulation: Current status and future prospects,

    J. J. Gross, “Emotion regulation: Current status and future prospects,” Psychological inquiry, vol. 26, no. 1, pp. 1–26, 2015

    2015

  40. [53]

    Psychological resilience: An affect-regulation framework,

    A. S. Troy, E. C. Willroth, A. J. Shallcross, N. R. Giuliani, J. J. Gross, and I. B. Mauss, “Psychological resilience: An affect-regulation framework,” Annual review of psychology, vol. 74, pp. 547–576, 2023

    2023

  41. [54]

    Graph attention networks,

    P . Veli ˇckovi´c, G. Cucurull, A. Casanova, A. Romero, P . Lio, and Y. Bengio, “Graph attention networks,” arXiv preprint arXiv:1710.10903, 2017

    2017 arXiv

  42. [55]

    Dynamic key-value memory networks for knowledge tracing,

    J. Zhang, X. Shi, I. King, and D.-Y. Yeung, “Dynamic key-value memory networks for knowledge tracing,” in Proceedings of the 26th international conference on World Wide Web, 2017, pp. 765–774

    2017

  43. [56]

    Convolutional knowledge tracing: Modeling individualization in student learning process,

    S. Shen, Q. Liu, E. Chen, H. Wu, Z. Huang, W. Zhao, Y. Su, H. Ma, and S. Wang, “Convolutional knowledge tracing: Modeling individualization in student learning process,” in Proceedings of the 43rd International ACM SIGIR Conference on Research and Develop- ment in Information ...

    2020

  44. [57]

    Ekt: Exercise-aware knowledge tracing for student performance prediction,

    Q. Liu, Z. Huang, Y. Yin, E. Chen, H. Xiong, Y. Su, and G. Hu, “Ekt: Exercise-aware knowledge tracing for student performance prediction,” IEEE Transactions on Knowledge and Data Engineering , vol. 33, no. 1, pp. 100–115, 2019

    2019

  45. [58]

    Assessing student’s dynamic knowledge state by exploring the question difficulty effect,

    S. Shen, Z. Huang, Q. Liu, Y. Su, S. Wang, and E. Chen, “Assessing student’s dynamic knowledge state by exploring the question difficulty effect,” in Proceedings of the 45th International ACM SIGIR Conference on Research and Development in Information Retrieval , 2022, pp. 427–437

    2022

  46. [59]

    Multimodal big data affective analytics: A comprehensive survey using text, audio, visual and physiological signals,

    N. J. Shoumy, L.-M. Ang, K. P . Seng, D. M. Rahaman, and T. Zia, “Multimodal big data affective analytics: A comprehensive survey using text, audio, visual and physiological signals,” Journal of Network and Computer Applications, vol. 149, p. 102447, 2020

    2020

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.