Pith. sign in

REVIEW 3 major objections 7 minor 74 references

Ranking-Based At-Risk Student Prediction Using Federated Learning and Differential Features

T0 review · 3 major / 7 minor · reviewed 2026-08-15 · deepseek-v4-flash

Pith's one-line read Federated learning matches centralized at-risk rankings

desk verdict Useful FL-for-EDM study with real data and code, but the claimed cross-course generalization is untested because every hold-out course is a later offering of a training course. read the letter →

arxiv 2505.09287 v2 pith:DWZ6S2NT submitted 2025-05-14 cs.LG cs.CY

classification cs.LGcs.CY
keywords at-riskstudentpredictionfederatedlearningdifferentialfeaturesriskrankingearlyprivacypreservationanalyticsrank-basedevaluation
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

The paper tries to show that a school can train an at-risk student warning system without pooling sensitive learning logs in one place. It combines federated learning, where each course trains locally and only model updates are shared, with differential features, which replace each student's raw behavior vector with pairwise differences between students. On data from 1,136 students across 12 course-offerings, the federated model ranks at-risk students about as accurately as a model trained on all data centrally, and more accurately than the same model trained on raw features. The same model also detects at-risk students from logs covering only the first half of a course nearly as well as from full-course logs. The practical payoff would be that many institutions could jointly build a high-performing early-warning tool without exposing grades or reading behavior.

What carries the argument

The load-bearing mechanism is the differential feature transform applied inside each client: every student is represented not by absolute activity counts but by differences to every other student in the same client. This does two jobs at once: it multiplies a client's $n$ samples into $n(n-1)$ training pairs, and it strips out client-level offsets such as overall activity level or course pacing, so the server model sees relative behavior rather than raw distributions. The other named component is FedAvg, the aggregation rule that averages locally trained weights proportionally to each client's sample count; the paper uses it to combine course-level models without moving raw logs. On top of the pairwise scores, a regression readout sums each student's predicted differences to recover an individual risk score for ranking.

What would settle it

Train the same federated pipeline on courses drawn from two institutions with very different overall activity levels, then check whether the cross-institution risk ranking stays stable; if the ranking degrades despite differential features, the core transferability claim fails. A cheaper check is to artificially rescale all feature values in one client by a large constant and see whether the model still ranks that client's students correctly.

Watch

Extended reading notes

Core claim

The central claim is that replacing absolute feature vectors with pairwise differences makes federated learning viable for educational early-warning systems. For each client, the method forms $d_{ij}=v_i-v_j$ from student feature representations and $e_{ij}=g_i-g_j$ from scored grades, trains a regression network locally on these pairs, and aggregates weights on the server with FedAvg. At prediction time the network outputs pairwise difference scores $p_{ij}$, which are summed per student as $q_i=\sum_{j\ne i}p_{ij}$ and sorted ascending to produce a risk ranking. The authors report that this pipeline matches centralized training on Top-n precision, nDCG, and PR-AUC across five hold-out courses, beats the same federated pipeline without differential features, and keeps most of its accuracy using only half the lecture sessions. They frame the result as a step toward privacy-preserving, generalizable student-performance prediction.

Load-bearing premise

The method assumes that replacing each student's raw behavior numbers with pairwise differences between students removes enough course-to-course and institution-to-institution variation that a model trained on some courses ranks unseen courses correctly, an assumption the authors admit has not been explicitly validated.

Editorial extensions

If this is right

  • A federated model can be trained across courses or institutions without any party seeing another party's raw learning logs or grades, while keeping ranking accuracy close to centralized training.
  • Switching features from absolute values to pairwise differences improves at-risk ranking on all five hold-out courses, suggesting the transform is the source of the cross-course generalization.
  • The trained model can be applied after half the lectures with little loss in ranking quality, so instructors can receive early warnings from the same model rather than retraining for each stage.
  • Because the output is a ranking rather than a hard at-risk label, instructors can act on the top-n list, and students near the boundary remain visible instead of being classified away.

Reading between the lines

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

  • If the cross-client invariance of pairwise differences holds, the same trick should transfer to other learning-management platforms and log formats, since constant offsets or scales in feature counts cancel in differences; a testable extension is applying the pipeline to clickstream logs from other systems.
  • The $n(n-1)$ expansion means small clients benefit most from differential features, so a federated network of small courses might see larger gains than the 12-course setup here.
  • The ranking readout is scale-free in a useful sense: if the regression's pairwise scores shift uniformly across students, the ranking is unchanged, which is the paper's stated reason early prediction works; this also implies calibration of absolute scores is unnecessary for deployment.
  • A privacy caveat not explored in the paper: sharing model updates, not raw data, reduces but does not eliminate leakage risk, and pairwise differences do not add formal privacy guarantees such as differential privacy; combining the method with noise would be a natural follow-up.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 7 minor

Summary. This paper proposes a method for at-risk student prediction that combines federated learning (FedAvg) with 'differential features,' i.e., pairwise differences between student feature vectors and between grade targets. The model is a small neural network regressor that produces individual risk scores by aggregating pairwise difference predictions, which are then ranked. Using e-book log data from 1,136 students across 12 course-instances at one university, the authors train 12 federated clients and evaluate on 5 hold-out course-instances. The main reported findings are that federated learning matches centralized learning in Top-n precision, nDCG, and PR-AUC, that differential features improve over non-differential features, and that the model can detect at-risk students using only the first half of the lecture data. The authors provide the modeling code publicly.

Significance. If the claims hold, the paper would offer a useful empirical contribution to privacy-preserving educational data mining: it demonstrates on real data that federated learning can match centralized performance for ranking-based at-risk prediction, and that pairwise difference regression improves ranking metrics. The use of real e-book logs from multiple courses and years, the direct comparison between federated and centralized training, and the release of code are clear strengths. However, the evidence as presented does not yet support the stronger claims of cross-course generalizability or a pre-specified early-prediction capability, because the hold-out courses are later offerings of seen course topics and the early-prediction cutoffs are chosen after inspecting test-set curves. The absence of any measure of variability over the 10 training runs further limits confidence in the reported improvements. These issues are fixable with additional experiments and re-framing, and the core FL-versus-centralized comparison is a solid starting point.

major comments (3)
  1. [§4.1.2, Tables 2–3, and §2.1] The hold-out test sets in Table 3 are later offerings of the same course topics as the training clients in Table 2: A-2022 follows A-2019/2020/2021, B-2020 follows B-2019, C-2022-1 follows C-2021-1/2, D-2022 follows D-2020/2021, and E-2021 follows E-2020-1/2. Courses F and G appear only in training, and no test course has a topic unseen in training. Consequently, the evaluation supports temporal holdout within known course content, not the cross-course or 'different course contents' generalization claimed in §2.1 and the abstract. Please either add a leave-one-course-out experiment (e.g., withhold all offerings of one course topic) or soften the generalizability claim to temporal generalization within the same course topics.
  2. [§4.2.3 and Table 6] The early-prediction cutoffs (4th lecture for A-2022 and B-2020; 8th for C-2022-1, D-2022, and E-2021) are selected after inspecting the PR-AUC curves in Figures 10–14, which are computed on the same test data used to report Table 6. Using test data to choose the evaluation point makes the reported early-prediction numbers in-sample and likely optimistic. Please pre-specify the cutoff (e.g., half of the lectures) before evaluation, or use a validation set to select the cutoff and report test performance only for that fixed choice.
  3. [§4.1.3, Tables 4 and 5] The paper reports averages over 10 training runs without any measure of dispersion (standard deviation, confidence interval) or significance test. Several proposed-versus-baseline differences are small (e.g., D-2022 PR-AUC 0.83 vs 0.82; A-2022 nDCG 0.83 vs 0.80; C-2022-1 n=15 Top-n precision 0.85 vs 0.79 for the non-differential baseline), so the claim that the proposed method 'consistently outperformed' the baseline is not supported without knowing run-to-run variability. At minimum, report the standard deviation across the 10 runs and, where possible, a paired test across runs.
minor comments (7)
  1. [§3.1.4 and Figure 4] The omission of L2 normalization is motivated by preserving action counts, but the effect on the distribution of feature magnitudes across clients is not discussed; this is directly relevant to the differential-features rationale in §3.2.1.
  2. [Equation (2)] Equation (2) defines Gm as a cumulative proportion (MaxScore times the cumulative count up to grade m), so the regression target is essentially a rank-based transformation of grades; this should be stated explicitly, because calling it a 'score' may mislead readers into thinking it is a standard grade point average.
  3. [General] Please correct typos: 'Predction' in Figure 2, 'substaintial' in §3.2.1, 'grads' in §4.1.2, and inconsistent capitalization of 'differential features' in headings.
  4. [§4.3.5] The paper candidly states that the assumption behind differential features has not been validated for a general setting. To strengthen the paper, add a quantitative check of whether pairwise differences reduce inter-client distribution distance (e.g., MMD or KL divergence) on the feature representations.
  5. [Table 3] The statement that the number of no-risk students (264) is 'more than twice' the number of at-risk students (127) is technically true but borderline (ratio 2.08); consider saying 'approximately twice.'
  6. [§4.2.1 and Abstract] The phrase 'more strongly preserved privacy' in §4.2.1 is appropriate, but the abstract's 'addresses privacy concerns' is stronger than what is demonstrated; FL without differential privacy or secure aggregation does not guarantee privacy against inference attacks. Please adjust the wording.
  7. [§4.1.2] In Section 4.1.2, the sentence 'The threshold of 15 was chosen to ensure that the number of at-risk students is sufficient for evaluating Top-n precision (n = 15)' is clear, but the definition of at-risk based on the 15th-ranked student in each course means the at-risk count varies across courses; consider reporting the sensitivity of the results to this threshold.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the empirical prediction pipeline is self-contained and independent of its evaluation labels.

full rationale

The derivation chain here is an empirical pipeline rather than a mathematical derivation: input features are 100-dimensional E2Vec representations of BookRoll e-book logs (Section 3.1.4), and labels are grade points mapped through Eq. 2 from final course grades (Section 3.1.5). These are measured independently of the model. Differential features (Eqs. 3-4) are a fixed data transformation of those inputs and labels, and the individual risk score qi is the sum of pairwise predictions (Eq. 5); no fitted parameter is renamed as a prediction, and the at-risk threshold (bottom 15) is used only for evaluation, with the paper explicitly stating that the boundary 'is set arbitrarily for evaluation and is not used to differentiate between at-risk and no-risk during model training.' Federated learning is FedAvg (Eq. 1), an external aggregation algorithm, and the comparison against centralized training on the same data is not forced by construction. The only cited work with overlapping authorship, E2Vec [35], is used as a feature extractor and is explicitly modified (L2-normalization omitted); it is not invoked as a uniqueness theorem or to rule out alternatives. Section 4.3.5 appropriately flags that the differential-feature distributional assumption 'has not yet been explicitly validated for a general setting,' and the hold-out test courses in Table 3 are temporally later offerings of courses present in the training data of Table 2, which weakens the cross-course generalization claim; these are validity and limitation concerns, not circularity. The central empirical comparisons therefore stand or fall on external data, not on self-referential reasoning.

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

The paper introduces no new entities. It relies on several free parameters and assumptions, most notably the grade-scaling constant, the at-risk definition, the post hoc early-prediction cutoffs, and the unvalidated differential-feature transferability assumption.

free parameters (4)
  • MaxScore = 0.95
    In Equation 2, grade scores are scaled by MaxScore=0.95, chosen ad hoc based on the convention that A corresponds to 90-100%. It directly sets the target scale for the regression.
  • At-risk threshold = 15
    Students ranked 15th or lower in actual final grades are labeled at-risk in evaluation (Section 4.1.2). The threshold is arbitrary and chosen to ensure enough students for Top-15 precision.
  • Early-prediction lecture cutoffs = 4 for A-2022, B-2020; 8 for C-2022-1, D-2022, E-2021
    Selected post hoc in Section 4.2.3 after examining test-set performance curves, not predefined.
  • Neural network hyperparameters = hidden sizes 50/10, dropout 0.2; learning rate, epochs, rounds not stated
    Architecture fixed without reported tuning; several training hyperparameters are omitted, making it hard to know how sensitive results are.
assumptions (5)
  • standard math FedAvg weighted averaging (Equation 1) produces a usable global model from heterogeneous local updates
    Relies on the standard federated averaging assumption; no convergence analysis is provided for these specific data.
  • domain assumption E2Vec features capture behavior relevant to final grades, even without L2 normalization
    The paper uses E2Vec embeddings from prior work [35] and deliberately omits normalization; this choice is not validated here.
  • domain assumption Federated learning alone preserves student privacy
    No differential privacy or secure aggregation is used; the paper equates 'no raw data transfer' with privacy, which is not a formal guarantee.
  • domain assumption Differential features mitigate inter-client feature distribution discrepancies
    This is the mechanism claimed for improved generalizability, but Section 4.3.5 concedes it is not explicitly validated.
  • domain assumption Cumulative grade scoring (Equation 2) is a valid transferable regression target
    Converting letter grades to client-specific cumulative proportions is an arbitrary choice that affects the target distribution and may not generalize.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Ranking-Based At-Risk Student Prediction Using Federated Learning and Differential Features." pith.science (2026). https://pith.science/paper/DWZ6S2NT

@misc{pith2026250509287,
  author       = {Pith},
  title        = {Pith review of: Ranking-Based At-Risk Student Prediction Using Federated Learning and Differential Features},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DWZ6S2NT}},
  note         = {Machine review of arXiv:2505.09287}
}
read the original abstract

Digital textbooks are widely used in various educational contexts, such as university courses and online lectures. Such textbooks yield learning log data that have been used in numerous educational data mining (EDM) studies for student behavior analysis and performance prediction. However, these studies have faced challenges in integrating confidential data, such as academic records and learning logs, across schools due to privacy concerns. Consequently, analyses are often conducted with data limited to a single school, which makes developing high-performing and generalizable models difficult. This study proposes a method that combines federated learning and differential features to address these issues. Federated learning enables model training without centralizing data, thereby preserving student privacy. Differential features, which utilize relative values instead of absolute values, enhance model performance and generalizability. To evaluate the proposed method, a model for predicting at-risk students was trained using data from 1,136 students across 12 courses conducted over 4 years, and validated on hold-out test data from 5 other courses. Experimental results demonstrated that the proposed method addresses privacy concerns while achieving performance comparable to that of models trained via centralized learning in terms of Top-n precision, nDCG, and PR-AUC. Furthermore, using differential features improved prediction performance across all evaluation datasets compared to non-differential approaches. The trained models were also applicable for early prediction, achieving high performance in detecting at-risk students in earlier stages of the semester within the validation datasets.

Figures

Figures reproduced from arXiv: 2505.09287 by the authors.

Figure 2
Figure 2. The sequence diagram of learning and prediction in [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 1
Figure 1. The overall framework of our proposed method [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 3
Figure 3. Overview of client-side learning 3.1.3 Learning Log Data in Digital Textbooks Digital textbook systems are educational platforms that al￾low learners to access and interact with learning materials through personal devices. While viewing the materials, learn￾ers can perform various actions, such as navigating forward or backward through pages, adding notes, or using markers. These interactions are recorded as learnin… view at source ↗
Figures from the paper (8 more)
Figure 5
Figure 5. Figure 5: Ranking creation method using a regression model [PITH_FULL_IMAGE:figures/full_fig_p004_5.png]
Figure 7
Figure 7. Figure 7: Use of relative values in differential features (the [PITH_FULL_IMAGE:figures/full_fig_p005_7.png]
Figure 6
Figure 6. Figure 6: Example of differential feature creation [PITH_FULL_IMAGE:figures/full_fig_p005_6.png]
Figure 9
Figure 9. Figure 9: Format of the learning logs in the BookRoll system [PITH_FULL_IMAGE:figures/full_fig_p006_9.png]
Figure 10
Figure 10. Figure 10: Relationship between lecture sessions and PR-AUC in course A-2022 1 2 3 4 5 6 7 Lecture 0.0 0.2 0.4 0.6 0.8 1.0 PR-AUC Proposed method Centralized + Difference Features Federated + No Difference Features Random [PITH_FULL_IMAGE:figures/full_fig_p009_10.png]
Figure 13
Figure 13. Figure 13: Relationship between lecture sessions and PR-AUC in course D-2022 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Lecture 0.0 0.2 0.4 0.6 0.8 1.0 PR-AUC Proposed method Centralized + Difference Features Federated + No Difference Features Random [PITH_FULL_IMAGE:figures/full_f…
Figure 17
Figure 17. Figure 17: Relationship of grades and rankings in early prediction for C-2022-1 F D C B A Grade 0 20 40 60 80 Rank [PITH_FULL_IMAGE:figures/full_fig_p010_17.png]
Figure 18
Figure 18. Figure 18: Relationship of grades and rankings in early prediction for D-2022 F D C B A Grade 0 10 20 30 40 50 Rank [PITH_FULL_IMAGE:figures/full_fig_p010_18.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

74 extracted references · 50 canonical work pages

  1. [1]

    Ranking-Based At-Risk Student Prediction Using Federated Learning and Differential Features

    INTRODUCTION Digital textbooks are widely used due to their capability to not only allow students to access learning materials on per- sonal devices but also collect records of their interactions as learning logs. These digital textbooks are now implemented in many educational institutions [43, 13, 8, 7], leading to the accumulation of vast amounts of lea...

  2. [2]

    higher-score students

    RELA TED WORK We aim to develop a generalized, high-performance ML model for predicting grades while preserving privacy using federated learning with differential features. Therefore, this section reviews prior studies related to this research across two themes: grade prediction and federated learning. 2.1 Prediction of Students’ Grades Predicting grades ...

  3. [3]

    risk ranking

    THE PROPOSED MODELING METHOD This section explains our method for creating a model that preserves privacy through federated learning and leverages differential features. The overall framework is illustrated in Figure 1. The following subsections focus on federated learn- ing and differential features, detailing how these approaches are used to conduct lea...

  4. [4]

    By employing a regression model, the extent of these differences can be explicitly learned, allowing the model to capture and utilize them

    Compatibility with Differential Features As described further in Section 3.2, the use of differ- ential features enables to establish higher/lower rela- tionships between students’ grades. By employing a regression model, the extent of these differences can be explicitly learned, allowing the model to capture and utilize them

  5. [5]

    However, this complicates identifying no-risk students who are close to being at-risk or at-risk students who are closer to no-risk

    Limitations of Classification Models Classification models divide students into at-risk and no-risk groups. However, this complicates identifying no-risk students who are close to being at-risk or at-risk students who are closer to no-risk. This limitation has been observed in previous EDM studies [49]. In con- trast, this study utilizes a regression mode...

  6. [6]

    Improved Generalization of the Predictions The proposed method enables to create highly general- izable models for prediction. For example, consider a scenario where a course consists of 16 lecture sessions, and the goal is to detect at-risk students based on the data available until the 6th lecture for early prediction. If a model trained on data from al...

  7. [7]

    Increase in Training Data If a client holds data for n students, the use of differ- ential features expands the number of data points to n(n− 1). This expansion helps mitigate overfitting and bias when training the local model on clients with lim- ited data because the increased data volume provides a richer dataset for training

  8. [8]

    pairwise difference scores

    Improved Generalization by Utilizing Relative Values Figure 7 illustrates that introducing differential fea- tures enables the use of relative values. Without dif- ferential features, absolute feature values are used for training. This can bias the server model because of differences in feature distributions among clients. For instance, clients with diffe...

Show all 74 references
  1. [9]

    operation name

    EXPERIMENTAL EV ALUA TION This section presents the comparison between the proposed method and baseline methods (defaults without the exper- imental condition). Additionally, it discusses the proposed method’s early prediction capabilities and examines the cor- responding resu...

  2. [10]

    In this study, we evaluate Top-n precision with four different settings: n ={5, 10, 15, At-risk}

    Top-n Precision Top-n precision indicates the proportion of students who are actually at-risk among the top-n students pre- dicted to incur the highest risk. In this study, we evaluate Top-n precision with four different settings: n ={5, 10, 15, At-risk}. Here, the “At-risk” r...

  3. [11]

    To compute nDCG, each student must be assigned a value

    Normalized Discounted Cumulated Gain (nDCG) As described in the reference paper [20], nDCG is a ranking-specific metric used to compare the predicted ranking order based on prediction values with the ac- tual ranking order based on the students’ grades. To compute nDCG, each s...

  4. [12]

    It plots Top-n precision (vertical axis) against Top-n recall (horizontal axis) as n varies, illus- trating their relationship

    Area Under the Precision-Recall Curve (PR-AUC) The Precision-Recall (PR) curve is commonly used in EDM research [17, 26] to evaluate predictive model performance. It plots Top-n precision (vertical axis) against Top-n recall (horizontal axis) as n varies, illus- trating their ...

  5. [13]

    Federated Learning

    Performance comparison between “Federated Learning” and “Centralized Machine Learning” Centralized ML is the baseline method in which the data from all clients are collected in a single location, with potential privacy concerns. The model is trained using the aggregated data

  6. [14]

    With Differential Features

    Performance comparison between “With Differential Features” and “Without Differential Features” Without differential features, applying the regression model to the feature representations yields individual prediction values for each student. As described in Section 3.1.6, stud...

  7. [15]

    CONCLUSION AND FUTURE WORK This study proposed a method that (1) applies federated learning in EDM to enable privacy-preserving prediction modeling and (2) leverages differential features to use rela- tive values between clients, resulting in a generalizable and high-performin...

  8. [16]

    Developing a More Generalizable Model As discussed in Section 4.3.1, the proposed approach is currently designed for learning log data from BookRoll. To develop a more generalizable and robust model, fur- ther work could investigate its applicability to different types of lear...

  9. [17]

    Alternative Integration Methods in Federated Learning This study employed the commonly used method called FedAvg for federated parameter integration. While FedAvg is a widely adopted baseline, other integration methods such as FedOpt [40] and SCAFFOLD [23] may offer advantages...

  10. [18]

    Identifying Actions Influencing Risk Levels Since the risk rankings are generated based on the learn- ing logs, further investigation is required to identify which specific student actions contribute to lower or higher risk, as discussed in Section 4.3.3. 5.2 A vailability of ...

  11. [19]

    ACKNOWLEDGMENTS This work was supported by JST CREST Grant Number JP- MJCR22D1 and JSPS KAKENHI Grant Number JP22H00551, Japan

  12. [20]

    Abdulraheem, R

    A. Abdulraheem, R. Abdullah Arshah, and H. Qin. Evaluating the effect of dataset size on predictive model using supervised learning technique. International Journal of Software Engineering & Computer Sciences (IJSECS) , 1(1):75–84, 02 2015. https://doi.org/10.15282/ijsecs.1.20...

  13. [21]

    Almoubayyed, S

    H. Almoubayyed, S. Fancsali, and S. Ritter. Generalizing predictive models of reading ability in adaptive mathematics software. In M. Feng, T. K ¨aser, and P. Talukdar, editors, Proceedings of the 16th International Conference on Educational Data Mining , pages 207–216, Bengal...

  14. [22]

    Altabrawee, O

    H. Altabrawee, O. Ali, and S. Qaisar. Predicting students’ performance using machine learning techniques. Journal of University of Babylon for pure and applied sciences, 27(1):194–205, 04 2019. https://doi.org/10.29196/jubpas.v27i1.2108

  15. [23]

    R. Baker. Getting past the current trade-off between privacy and equity in educational technology. The Economics of Equity in K-12 Education: Connecting Financial Investments with Effective Programming , 123, 2023. https://scholar.google.com/citations?view_op= view_citation&ci...

  16. [24]

    R. S. Baker, S. Hutt, C. A. Brooks, N. Srivastava, and C. Mills. Open science and educational data mining: Which practices matter most? In B. Paaßen and C. D. Epp, editors, Proceedings of the 17th International Conference on Educational Data Mining , pages 279–287, Atlanta, Ge...

  17. [25]

    M. K. Belaid, M. Rabus, and E. H ¨ullermeier. Pairwise difference learning for classification. In D. Pedreschi, A. Monreale, R. Guidotti, R. Pellungrini, and F. Naretto, editors, Discovery Science - 27th International Conference, DS 2024, Pisa, Italy, October 14-16, 2024, Proc...

  18. [26]

    Boulanger and V

    D. Boulanger and V. Kumar. An overview of recent developments in intelligent e-textbooks and reading analytics. In S. A. Sosnovsky, P. Brusilovsky, R. G. Baraniuk, R. Agrawal, and A. S. Lan, editors, iTextbooks@AIED, volume 2384, pages 44–56. CEUR-WS.org, 2019. https: //api.se...

  19. [27]

    Brusilovsky, S

    P. Brusilovsky, S. Sosnovsky, and K. Thaker. The return of intelligent textbooks. AI Magazine, 43(3):337–340, 8 2022. https://doi.org/10.1002/aaai.12061

  20. [28]

    C.-H. Chen, S. J. H. Yang, J.-X. Weng, H. Ogata, and C.-Y. Su. Predicting at-risk university students based on their e-book reading behaviours by using machine learning classifiers. Australasian Journal of Educational Technology, 37(4):130–144, Jun. 2021. https://doi.org/10.14...

  21. [29]

    J. M. Cock, M. Marras, C. Giang, and T. K ¨aser. Generalisable methods for early prediction in interactive simulations for education. In A. Mitrovic and N. Bosch, editors, Proceedings of the 15th International Conference on Educational Data Mining , volume abs/2207.01457, page...

  22. [30]

    Csurka, C

    G. Csurka, C. Dance, L. Fan, J. Willamowski, and C. Bray. Visual categorization with bags of keypoints. In Workshop on statistical learning in computer vision, ECCV, volume 1, pages 1–2. Prague, 2004. https://www.cse.unr.edu/~bebis/CS773C/ ObjectRecognition/Papers/Dance04.pdf

  23. [31]

    Delianidi, K

    M. Delianidi, K. I. Diamantaras, G. Chrysogonidis, and V. Nikiforidis. Student Performance Prediction Using Dynamic Neural Models. In S. I. Hsiao, S. S. Sahebi, F. Bouchet, and J. Vie, editors, Proceedings of the 14th International Conference on Educational Data Mining , volum...

  24. [32]

    A. M. Embong, A. M. Noor, H. M. Hashim, R. M. Ali, and Z. H. Shaari. E-books as textbooks in the classroom. Procedia - Social and Behavioral Sciences , 47:1802–1809, 2012. https://doi.org/10.1016/j.sbspro.2012.06.903

  25. [33]

    Flanagan and H

    B. Flanagan and H. Ogata. Learning analytics platform in higher education in japan. Knowledge Management and E-Learning, 10:469–484, 11 2018. https://doi.org/10.34105/j.kmel.2018.10.029

  26. [34]

    L. Gao, H. Fu, L. Li, Y. Chen, M. Xu, and C.-Z. Xu. Feddc: Federated learning with non-iid data via local drift decoupling and correction. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 10112–10121. IEEE, June 2022. https://d...

  27. [35]

    A. Haim, R. Gyurcsan, C. Baxter, S. T. Shaw, and N. T. Heffernan. How to Open Science: Debugging Reproducibility within the Educational Data Mining Conference. In M. Feng, T. K ¨aser, and P. Talukdar, editors, Proceedings of the 16th International Conference on Educational Dat...

  28. [36]

    Hlosta, Z

    M. Hlosta, Z. Zdrahal, and J. Zendulka. Ouroboros: early identification of at-risk students without models based on legacy data. In M. Hatala, A. Wis, P. Winne, G. Lynch, X. Ochoa, I. Molenaar, S. Dawson, S. Shehata, and J. P.-L. Tan, editors, Proceedings of the Seventh Intern...

  29. [37]

    M. Hoq, P. Brusilovsky, and B. Akram. Analysis of an explainable student performance prediction model in an introductory programming course. In M. Feng, T. K¨aser, and P. Talukdar, editors, Proceedings of the 16th International Conference on Educational Data Mining, pages 79–9...

  30. [38]

    S. Hutt, R. S. Baker, M. M. Ashenafi, J. M. Andres-Bray, and C. Brooks. Controlled outputs, full data: A privacy-protecting infrastructure for mooc data. British Journal of Educational Technology , 53(4):756–775, 5 2022. https://doi.org/10.1111/bjet.13231

  31. [39]

    J ¨arvelin and J

    K. J ¨arvelin and J. Kek¨al¨ainen. Cumulated gain-based evaluation of ir techniques. ACM Transactions on Information Systems (TOIS) , 20(4):422–446, 10 2002. https://doi.org/10.1145/582415.582418

  32. [40]

    Kairouz, H

    P. Kairouz, H. B. McMahan, B. Avent, A. Bellet, M. Bennis, A. Nitin Bhagoji, K. Bonawitz, Z. Charles, G. Cormode, R. Cummings, R. G. L. D’Oliveira, H. Eichner, S. El Rouayheb, D. Evans, J. Gardner, Z. Garrett, A. Gasc´ on, B. Ghazi, P. B. Gibbons, M. Gruteser, Z. Harchaoui, C....

  33. [41]

    Kanamori, T

    S. Kanamori, T. Abe, T. Ito, K. Emura, L. Wang, S. Yamamoto, T. P. Le, K. Abe, S. Kim, R. Nojima, et al. Privacy-preserving federated learning for detecting fraudulent financial transactions in japanese banks. Journal of Information Processing , 30(0):789–795, 2022. https://do...

  34. [42]

    S. P. Karimireddy, S. Kale, M. Mohri, S. J. Reddi, S. U. Stich, and A. T. Suresh. SCAFFOLD: stochastic controlled averaging for on-device federated learning. CoRR, abs/1910.06378, 10 2019. http://arxiv.org/abs/1910.06378

  35. [43]

    Keskin and H

    S. Keskin and H. Yurdug ¨ul. E-learning experience: Modeling students’ e-learning interactions using log data. Journal of Educational Technology and Online Learning, 5(1):1–13, 01 2022. https://doi.org/10.31681/jetol.938363

  36. [44]

    Klose, V

    M. Klose, V. Desai, Y. Song, and E. F. Gehringer. Edm and privacy: Ethics and legalities of data collection, usage, and storage. In A. N. Rafferty, J. Whitehill, C. Romero, and V. Cavalli-Sforza, editors, Educational Data Mining. International Educational Data Mining Society, ...

  37. [45]

    Koutcheme, S

    C. Koutcheme, S. Sarsa, A. Hellas, L. Haaranen, and J. Leinonen. Methodological considerations for predicting at-risk students. In J. Sheard and P. Denny, editors, Proceedings of the 24th Australasian Computing Education Conference, pages 105–113. ACM, 02 2022. https://doi.org...

  38. [46]

    Leelaluk, C

    S. Leelaluk, C. Tang, V. ˇSv´ abensk´ y, and A. Shimada. Knowledge distillation in rnn-attention models for early prediction of student performance. arXiv preprint arXiv:2412.14526, abs/2412.14526, 12 2024. https://doi.org/10.48550/arxiv.2412.14526

  39. [47]

    J. Li, S. Supraja, W. Qiu, and A. W. H. Khong. Grade prediction via prior grades and text mining on course descriptions: Course outlines and intended learning outcomes. In A. Mitrovic and N. Bosch, editors, Proceedings of the 15th International Conference on Educational Data M...

  40. [48]

    L. Li, Y. Fan, M. Tse, and K.-Y. Lin. A review of applications in federated learning. Computers & Industrial Engineering, 149:106854, 11 2020. https://doi.org/10.1016/j.cie.2020.106854

  41. [49]

    Q. Li, Y. Diao, Q. Chen, and B. He. Federated learning on non-iid data silos: An experimental study. In 2022 IEEE 38th International Conference on Data Engineering (ICDE), pages 965–978. IEEE, 5 2022. https://doi.org/10.1109/ICDE53745.2022.00077

  42. [50]

    X. Li, M. Jiang, X. Zhang, M. Kamp, and Q. Dou. Fedbn: Federated learning on non-iid features via local batch normalization. arXiv preprint arXiv:2102.07623 , abs/2102.07623, 2 2021. https://arxiv.org/pdf/2102.07623

  43. [51]

    Y. Liu, J. James, J. Kang, D. Niyato, and S. Zhang. Privacy-preserving traffic flow prediction: A federated learning approach. IEEE Internet of Things Journal , 7(8):7751–7763, 8 2020. https://doi.org/10.1109/jiot.2020.2991401

  44. [52]

    Z. Liu, X. Jiao, C. Li, and W. Xing. Fair prediction of students’ summative performance changes using online learning behavior data. In B. Paaßen and C. D. Epp, editors, Proceedings of the 17th International Conference on Educational Data Mining , pages 686–691, Atlanta, Georg...

  45. [53]

    McMahan, E

    B. McMahan, E. Moore, D. Ramage, S. Hampson, and B. A. y Arcas. Communication-efficient learning of deep networks from decentralized data. In A. Singh and X. J. Zhu, editors, Artificial intelligence and statistics , volume 54, pages 1273–1282. PMLR, JMLR.org, 4 2017. https://p...

  46. [54]

    Miyazaki, V

    Y. Miyazaki, V. ˇSv´ abensk´ y, Y. Taniguchi, F. Okubo, T. Minematsu, and A. Shimada. E2vec: Feature embedding with temporal information for analyzing student actions in e-book systems. In B. Paaßen and C. D. Epp, editors, Proceedings of the 17th International Conference on Ed...

  47. [55]

    Ogata, C

    H. Ogata, C. Yin, M. Oi, F. Okubo, A. Shimada, K. Kojima, and M. Yamada. e-book-based learning analytics in university education. In Doctoral Student Consortium (DSC) - Proceedings of the 23rd International Conference on Computers in Education, ICCE 2015, pages 401–406. Asia-P...

  48. [56]

    Oldenhof, G

    M. Oldenhof, G. ´Acs, B. Pej´ o, A. Schuffenhauer, N. Holway, N. Sturm, A. Dieckmann, O. Fortmeier, E. Boniface, C. Mayer, et al. Industry-scale orchestrated federated learning for drug discovery. In B. Williams, Y. Chen, and J. Neville, editors, Proceedings of the AAAI Confer...

  49. [58]

    N. Ong, J. Zhu, and D. Mosse. Towards including instructor features in student grade prediction. In A. Mitrovic and N. Bosch, editors, Proceedings of the 15th International Conference on Educational Data Mining, pages 239–250, Durham, United Kingdom, July

  50. [60]

    Radmehr, A

    B. Radmehr, A. Singla, and T. K ¨aser. Towards generalizable agents in text-based educational environments: A study of integrating rl with llms. In B. Paaßen and C. D. Epp, editors, Proceedings of the 17th International Conference on Educational Data Mining, volume abs/2404.18...

  51. [61]

    S. J. Reddi, Z. Charles, M. Zaheer, Z. Garrett, K. Rush, J. Koneˇ cn´ y, S. Kumar, and H. B. McMahan. Adaptive federated optimization. CoRR, abs/2003.00295, 2 2020. https://doi.org/10.48550/arxiv.2003.00295

  52. [62]

    Rohani, K

    N. Rohani, K. Gal, M. Gallagher, and A. Manataki. Early prediction of student performance in a health data science mooc. In M. Feng, T. K ¨aser, and P. Talukdar, editors, Proceedings of the 16th International Conference on Educational Data Mining , pages 325–333, Bengaluru, In...

  53. [63]

    Shimada and S

    A. Shimada and S. Konomi. A lecture supporting system based on real-time learning analytics. In 14th International Conference On Cognition And Exploratory Learning In The Digital Age (CELDA 2017), pages 197–204, 10 2017. https: //files.eric.ed.gov/fulltext/ED579463.pdf

  54. [64]

    J. Staiger. How e-books are used: A literature review of the e-book studies conducted from 2006 to 2011. Reference and User Services Quarterly , 51(4):355–365, 6 2012. https://doi.org/10.5860/rusq.51n4.355

  55. [65]

    C. Sun, A. Shrivastava, S. Singh, and A. Gupta. Revisiting unreasonable effectiveness of data in deep learning era. In 2017 IEEE International Conference on Computer Vision (ICCV) , pages 843–852. IEEE, 10

  56. [67]

    Trask, D

    T. Trask, D. N. Lytle, M. Boyle, D. D. Joyner, and D. A. Mubarak. A comparative analysis of student performance predictions in online courses using heterogeneous knowledge graphs. In B. Paaßen and C. D. Epp, editors, Proceedings of the 17th International Conference on Educatio...

  57. [68]

    Tynes, W

    M. Tynes, W. Gao, D. J. Burrill, E. R. Batista, D. Perez, P. Yang, and N. Lubbers. Pairwise difference regression: A machine learning meta-algorithm for improved prediction and uncertainty quantification in chemical search. Journal of Chemical Information and Modeling, 61(8):3...

  58. [69]

    van Haastrecht, M

    M. van Haastrecht, M. Brinkhuis, and M. Spruit. Federated learning analytics: Investigating the privacy-performance trade-off in machine learning for educational analytics. In A. M. Olney, I.-A. Chounta, Z. Liu, O. C. Santos, and I. I. Bittencourt, editors, International Confe...

  59. [70]

    ˇSv´ abensk´ y, R

    V. ˇSv´ abensk´ y, R. Baker, A. Zambrano, Y. Zou, and S. Slater. Towards generalizable detection of urgency of discussion forum posts. In M. Feng, T. K ¨aser, and P. Talukdar, editors, Proceedings of the 16th International Conference on Educational Data Mining , volume abs/230...

  60. [71]

    ˇSv´ abensk´ y, K

    V. ˇSv´ abensk´ y, K. Tk´ aˇ cik, A. Birdwell, R. Weiss, R. S. Baker, P. ˇCeleda, J. Vykopal, J. Mache, and A. Chattopadhyay. Detecting Unsuccessful Students in Cybersecurity Exercises in Two Different Learning Environments. In Proceedings of the 54th Frontiers in Education Co...

  61. [72]

    K. D. Wang, J. M. Cock, T. K ¨aser, and E. Bumbacher. A systematic review of empirical studies using log data from open-ended learning environments to measure science and engineering practices. British Journal of Educational Technology, 54(1):192–221, 1 2023. https://doi.org/1...

  62. [73]

    Zheng, H

    Q. Zheng, H. He, T. Ma, N. Xue, B. Li, and B. Dong. Big log analysis for e-learning ecosystem. In Y. Li, X. Fei, K.-M. Chao, and J.-Y. Chung, editors, 2014 IEEE 11th International Conference on e-Business Engineering, pages 258–263. IEEE, 11 2014. https://doi.org/10.1109/ICEBE.2014.51

  63. [74]

    H. Zhu, J. Xu, S. Liu, and Y. Jin. Federated learning on non-iid data: A survey. Neurocomputing, 465:371–390, 9 2021. https://doi.org/10.1016/j.neucom.2021.07.098

  64. [2017]

    https://doi.org/10.1109/ICCV.2017.97

  65. [2022]

    https://doi.org/10.5281/zenodo.6853063

    International Educational Data Mining Society. https://doi.org/10.5281/zenodo.6853063

  66. [2023]

    https://doi.org/10.1609/aaai.v37i13.26847

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.