Pith. sign in

REVIEW 4 major objections 7 minor 39 references

Representation Learning on Out of Distribution in Tabular Data

T0 review · 4 major / 7 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read A CPU-trained contrastive model, TCL, is reported to beat GPU-heavy deep networks on out-of-distribution tabular classification while staying competitive on regression.

desk verdict TCL is a plausible lightweight tabular contrastive learner, but the OOD comparison is built on hand-picked splits and a broken baseline, so the headline claim is not supported. read the letter →

arxiv 2502.10095 v2 pith:N6GGVXZB submitted 2025-02-14 cs.LG

Achmad Ginanjar , Xue Li , Priyanka Singh , Wen Hua This is my paper · ORCID
classification cs.LG
keywords out-of-distributionpredictiontabulardatacontrastivelearningrepresentationCPUtrainingefficiency-accuracytrade-offOpenMaxtemperaturescaling
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 argues that a lightweight contrastive-learning model for tabular data, called TCL, can handle out-of-distribution (OOD) prediction better than much heavier deep models while running entirely on ordinary CPUs. The authors build OOD test sets by applying OpenMax and TemperatureScaling to separate low-confidence rows from each of ten tabular datasets, then compare twelve baselines including FT-Transformer and ResNet. They report that TCL achieves the best F1 on most classification datasets, competitive RMSE on regression datasets, and the highest speed/accuracy trade-off because its training time is far shorter. If true, this means strong OOD tabular performance is achievable without GPU access.

What carries the argument

TCL (Tabular Contrastive Learning) is the central mechanism. The key design choice is full matrix augmentation: both augmented views contain the complete row of features instead of sliced subsets, and the simplified contrastive loss is a dot product between the two encoded views divided by a temperature, so the encoder is forced to make the representation of a row invariant to added noise. A reconstruction loss and a distance loss between encodings ensure that the representation is not degenerate.

What would settle it

Re-run the comparison with OOD test sets defined by a fixed, non-manual rule, for example holding out whole classes or taking the lowest-scoring 10% of rows by OpenMax score without inspecting model performance, and check whether TCL still leads on classification; if the margin shrinks, reverses, or the regression gap widens, the central claim is an artifact of the hand-picked thresholds.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that a contrastive representation learner operating on entire tabular rows, not feature subsets, can learn features that transfer to out-of-distribution test data better than the features learned by heavier architectures. TCL's training creates two noisy copies of each row, pushes their encoded forms together with a temperature-scaled dot-product loss, and reconstructs the input, while at inference only the encoder remains. Across ten datasets the authors report that TCL has the best F1 on most classification benchmarks, is close to the best on regression, and achieves the best speed/accuracy trade-off because its CPU training time is a small fraction of the GPU training time of FT-Transformer and ResNet.

Load-bearing premise

The entire comparison depends on the manually chosen thresholds from OpenMax and TemperatureScaling score plots; if those thresholds do not genuinely isolate out-of-distribution rows, the reported gap between TCL and the other models reflects the split rather than the model.

Editorial extensions

If this is right

  • If TCL's reported results hold, users with only CPU hardware can train a tabular model that outperforms GPU-trained FT-Transformer and ResNet on OOD classification tasks.
  • The short training time makes extensive hyperparameter search feasible on consumer hardware, which usually favours the model that can be tuned more thoroughly.
  • TCL's unsupervised contrastive step can use unlabeled rows ahead of the supervised head, so adding unlabeled data could strengthen the learned representation.
  • The paper's OOD detection recipe, OpenMax plus TemperatureScaling with manual thresholds, gives practitioners a direct way to construct OOD test sets from a single tabular dataset.

Reading between the lines

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

  • A natural next test is to feed TCL's learned representations into a tree-based classifier, since the paper only tests the encoder with its own supervised head.
  • The manual OOD threshold selection could be automated by choosing thresholds on a held-out calibration set; this would turn the paper's recipe into a standardised benchmark and reveal how sensitive the ranking is to the split.
  • Because the contrastive loss uses only dot-product similarity between two noisy views, adding corruptions beyond Gaussian noise, such as feature masking or swapping, might further improve OOD robustness.
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

4 major / 7 minor

Summary. The paper proposes TCL, a contrastive-learning representation method for tabular data, and claims that it outperforms existing deep models such as FT-Transformer and ResNet on out-of-distribution (OOD) classification while remaining competitive on regression, and that it does so with much lower training cost on CPU hardware. The experiments cover ten tabular datasets, use OpenMax and TemperatureScaling to split data into in-distribution and OOD sets, compare TCL with twelve baselines, and report F1/RMSE, training time, and a speed/accuracy trade-off metric. The paper also presents practical guidance for detecting and visualizing OOD data.

Significance. If the central claim were established, the contribution would be practically valuable: a CPU-trainable contrastive method that beats GPU-trained deep models on OOD tabular classification would lower the hardware barrier for many practitioners. The paper also makes a potentially useful effort to expose OOD detection and evaluation as an accessible workflow, and it provides a public code link. However, the evidence as presented does not support the headline claims. The OOD test sets are constructed from manually selected thresholds on model-based confidence scores without external ground truth for out-of-distribution samples, the main comparison table has no error bars or repeated runs, several baseline entries are missing, and at least one baseline (SCARF) reports F1=0.00 on multiple datasets, which suggests a broken implementation rather than a meaningful comparison. The significance is therefore conditional on a substantially stronger evaluation.

major comments (4)
  1. [Section VI.B, Tables II and III] The OOD test splits are load-bearing for the paper's central claim, but their validity is not established. The thresholds are 'manually assigned by observing the graphs' and validated only by the expectation that performance decreases on the separated tail. This is circular: any model with calibrated confidence will perform worse on low-confidence samples, so TCL's better Table III scores may reflect handling of low-confidence points rather than generalization to a genuinely different distribution pood. The paper needs an independent definition of OOD, such as a principled covariate-shift construction or an external OOD benchmark, plus a sensitivity analysis showing that small threshold perturbations do not change the ranking.
  2. [Table III] The main performance comparison is reported without error bars, standard deviations, or information about the number of seeds or runs. Many entries are missing (for example, FT-Transformer is absent for CoverType, Yahoo, and Microsoft, and several models are missing on multiple datasets), while SCARF obtains F1=0.00 on Helena and Aloi and near-zero values on other datasets. A zero F1 on a multi-class dataset is not a plausible trained-model result and suggests a broken or misconfigured baseline. Without complete baselines and repeated trials, the claim that 'TCL outperforms other models' in Table III is not supported.
  3. [Section VII.C and Table IV] The training-time comparison is not apples-to-apples. TCL was trained on a CPU while the deep baselines were trained on an H100 GPU, and the time recorded for TCL is the time to 'stabilise their loss' (around 15 epochs), whereas the times for other models appear to include extensive hyperparameter tuning, with Yahoo and Microsoft requiring five days of tuning. The subsequent speed/accuracy trade-off in Table V therefore mixes tuning time, final training time, and different hardware, so the efficiency advantage is not cleanly quantified. The paper should report a fixed training protocol with matched epochs/budgets and separate tuning time from final training time.
  4. [Section III.A, Eq. (5) and Definition 3] The efficiency-accuracy objective is not precisely defined. The text first writes T = P/t, then defines T as 'max P_ood' over 'min t', and the regression adjustment uses 1/P without a clear derivation. Because Table V is used to support a major contribution, this metric needs a formal definition that is applied consistently across classification and regression, including how the reported trade-off values in Table V are computed from the F1/RMSE and time values in Tables III and IV.
minor comments (7)
  1. [Throughout] The notation is inconsistent: 'RestNet' and 'Resnet' appear alongside 'ResNet', 'GDBT' is used for gradient boosting, and 'Scarfc' should be 'SCARF'. The paper would benefit from a careful proofreading pass.
  2. [Table II] The threshold column contains entries such as '1.46eE-03' and '-8.30eE-03', which appear to be typographical errors in scientific notation. The detection method used for each dataset is also only given as a one-letter code without an explanation in the table caption.
  3. [Section IV.A, Eq. (9)] The simplified contrastive loss formula 'MSE([0], dot(xe1 · xe2))/T' is malformed as written. It should be stated in standard notation with the batch dimension and the temperature parameter clearly defined.
  4. [Section VII.A, Table II] The validation of OOD separation using linear regression is not described. The reader is not told what regression is performed, which variables are used, or what result would confirm that the separated set is genuinely out-of-distribution.
  5. [Section VI.B] The paper does not provide the actual threshold values or the sample sizes M and N for each dataset. Since the thresholds are a free parameter of the evaluation, they should be reported in full for reproducibility.
  6. [Section V] The description of the augmentation and training procedure is incomplete: no noise distribution or magnitude is given, and the encoder/decoder architectures are not specified beyond 'one hidden layer and one normalisation layer'. This makes it difficult to reproduce TCL from the text alone.
  7. [Section VIII] The conclusion acknowledges limitations such as the need for continual learning and interpretability, but it does not acknowledge the more immediate issue that the OOD evaluation protocol itself is not validated. A discussion of the limitations of the manual thresholding approach would be appropriate.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: TCL's architecture and loss derivation are self-contained, and the cited CFL work is used only for method provenance, not as evidence for the reported OOD results.

full rationale

The paper's derivation chain is self-contained: Definition 2 defines OOD prediction via a distinct distribution pood, and the proposed TCL method (Section IV) is defined through full-matrix augmentation, an encoder-decoder structure, and explicit loss terms (Eqs. 6-9) that do not reference the OOD test split or any OOD-specific fitted parameter. The reported improvements in Tables III-V are empirical comparisons on datasets separated by manual thresholds in Section VI.B; while those thresholds are hand-picked and may affect what counts as 'OOD', they are not estimated from TCL's output and do not appear in TCL's objective, so the comparison is not forced by construction. The citation to the authors' CFL work [11] is explicitly described as a 'local adaptation' and is used for provenance of the contrastive-learning design, not as load-bearing evidence for TCL's performance; no uniqueness theorem or unverified prior result is imported to rule out alternatives. The statement in Section VI.B that the authors expected performance to decrease after OOD separation is a sanity check, not a fitted prediction. Thus no circular step meeting the quoted-evidence bar is present, and the central claim stands on independent experimental comparison, although the manual OOD threshold construction is a legitimate correctness/evaluation concern rather than a circularity one.

Assumptions & free parameters 1 free parameters · 3 assumptions · 0 invented entities

The main free parameter is the per-dataset OOD threshold used to build the test split. The method relies on the assumption that manual thresholding produces valid OOD boundaries and that full-matrix augmentation is beneficial. No invented entities are introduced.

free parameters (1)
  • Per-dataset OOD detection threshold = Manual selections in Table II, e.g., Adult 0.1628, Helena 0.045, Jannis -0.02, Higgs 0.042, Aloi 0.016, Covtype…
    Each threshold is chosen by visual inspection of score distributions (Section VI.B) and defines the in-distribution/OOD split used for training and testing. The split directly determines the reported F1 and RMSE values.
assumptions (3)
  • domain assumption OpenMax and TemperatureScaling produce valid out-of-distribution scores for tabular data
    Sections II.A and VI.B use these image-derived methods as the basis for OOD separation; if their scores do not rank tabular OOD correctly, the whole evaluation rests on invalid splits.
  • ad hoc to paper Manually selected thresholds yield a fair and meaningful OOD test set
    Section VI.B states thresholds were 'manually assigned by observing the graphs'; this introduces a human-chosen, dataset-specific boundary that is not validated by any held-out procedure.
  • domain assumption Full matrix augmentation preserves the information required for downstream prediction
    Section IV.A claims full matrix augmentation captures comprehensive feature interactions, but no ablation or theoretical argument supports this claim.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Representation Learning on Out of Distribution in Tabular Data." pith.science (2026). https://pith.science/paper/N6GGVXZB

@misc{pith2026250210095,
  author       = {Pith},
  title        = {Pith review of: Representation Learning on Out of Distribution in Tabular Data},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/N6GGVXZB}},
  note         = {Machine review of arXiv:2502.10095}
}
read the original abstract

The open-world assumption in model development suggests that a model might lack sufficient information to adequately handle data that is entirely distinct or out of distribution (OOD). While deep learning methods have shown promising results in handling OOD data through generalization techniques, they often require specialized hardware that may not be accessible to all users. We present TCL, a lightweight yet effective solution that operates efficiently on standard CPU hardware. Our approach adapts contrastive learning principles specifically for tabular data structures, incorporating full matrix augmentation and simplified loss calculation. Through comprehensive experiments across 10 diverse datasets, we demonstrate that TCL outperforms existing models, including FT-Transformer and ResNet, particularly in classification tasks, while maintaining competitive performance in regression problems. TCL achieves these results with significantly reduced computational requirements, making it accessible to users with limited hardware capabilities. This study also provides practical guidance for detecting and evaluating OOD data through straightforward experiments and visualizations. Our findings show that TCL offers a promising balance between performance and efficiency in handling OOD prediction tasks, which is particularly beneficial for general machine learning practitioners working with computational constraints.

Figures

Figures reproduced from arXiv: 2502.10095 by the authors.

Figure 1
Figure 1. Tabular Contrastive Learning (TCL). The data [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

39 extracted references · 35 canonical work pages

  1. [1]

    Generalized odin: Detecting out-of-distribution image without learning from out- of-distribution data,

    Y .-C. Hsu, Y . Shen, H. Jin, and Z. Kira, “Generalized odin: Detecting out-of-distribution image without learning from out- of-distribution data,” 2020

    work page 2020

  2. [2]

    A baseline for detecting mis- classified and out-of-distribution examples in neural networks,

    D. Hendrycks and K. Gimpel, “A baseline for detecting mis- classified and out-of-distribution examples in neural networks,” 5th International Conference on Learning Representations, ICLR 2017 - Conference Track Proceedings , 10 2016. TABLE II: The OOD detection settings. Performances are results of models trained with linear regression (r2) and logistic r...

    work page 2017

  3. [3]

    Computational power and the social impact of artificial intelligence,

    T. Hwang, “Computational power and the social impact of artificial intelligence,” SSRN Electronic Journal , 2018

    work page 2018

  4. [4]

    The de-democratization of ai: Deep learning and the compute divide in artificial intelligence research,

    N. Ahmed and M. Wahed, “The de-democratization of ai: Deep learning and the compute divide in artificial intelligence research,” 2020

    work page 2020

  5. [5]

    Multi-Class Data Description for Out-of-distribution Detection,

    D. Lee, S. Yu, and H. Yu, “Multi-Class Data Description for Out-of-distribution Detection,” Proceedings of the ACM SIGKDD International Conference on Knowledge Discovery and Data Mining , pp. 1362–1370, aug 2020

    work page 2020

  6. [6]

    Multi-class data description for out-of-distribution detection,

    D. Lee, S. Yu, and H. Yu, “Multi-class data description for out-of-distribution detection,” in Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, KDD ’20, (New York, NY , USA), p. 1362–1370, Association for Computing Machinery, 2020

    work page 2020

  7. [7]

    Towards open set deep net- works,

    A. Bendale and T. E. Boult, “Towards open set deep net- works,” Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition , vol. 2016- December, pp. 1563–1572, 12 2016

    work page 2016

  8. [8]

    Dropout as a bayesian approxi- mation: Representing model uncertainty in deep learning,

    Y . Gal and Z. Ghahramani, “Dropout as a bayesian approxi- mation: Representing model uncertainty in deep learning,” 6 2016

    work page 2016

Show all 39 references
  1. [9]

    Enhancing the reliability of out-of-distribution image detection in neural networks,

    S. Liang, Y . Li, and R. Srikant, “Enhancing the reliability of out-of-distribution image detection in neural networks,” 6th International Conference on Learning Representations, ICLR 2018 - Conference Track Proceedings , 6 2017

    2018

  2. [10]

    Why do tree- based models still outperform deep learning on typical tabular data?,

    L. Grinsztajn, E. Oyallon, and G. Varoquaux, “Why do tree- based models still outperform deep learning on typical tabular data?,” Advances in Neural Information Processing Systems , vol. 35, pp. 507–520, dec 2022

    2022

  3. [11]

    Contrastive federated learn- ing with tabular data silos,

    A. Ginanjar, X. Li, and W. Hua, “Contrastive federated learn- ing with tabular data silos,” 2024

    2024

  4. [12]

    A simple framework for contrastive learning of visual representations,

    T. Chen, S. Kornblith, M. Norouzi, and G. Hinton, “A simple framework for contrastive learning of visual representations,” pp. 1597–1607, 11 2020

    2020

  5. [13]

    Subtab: Subset- ting features of tabular data for self-supervised representation learning,

    T. Ucar, E. Hajiramezanali, and L. Edwards, “Subtab: Subset- ting features of tabular data for self-supervised representation learning,” vol. 23, 2021

    2021

  6. [14]

    Speed/accuracy trade-offs for modern convolutional object detectors,

    J. Huang, V . Rathod, C. Sun, M. Zhu, A. Korattikara, A. Fathi, I. Fischer, Z. Wojna, Y . Song, S. Guadarrama, and K. Murphy, “Speed/accuracy trade-offs for modern convolutional object detectors,” 2017

    2017

  7. [15]

    Probabilistic outputs for support vector ma- chines and comparisons to regularized likelihood methods,

    J. Platt et al. , “Probabilistic outputs for support vector ma- chines and comparisons to regularized likelihood methods,” Advances in large margin classifiers , vol. 10, no. 3, pp. 61– 74, 1999

    1999

  8. [16]

    On calibra- tion of modern neural networks,

    C. Guo, G. Pleiss, Y . Sun, and K. Q. Weinberger, “On calibra- tion of modern neural networks,” 34th International Confer- ence on Machine Learning, ICML 2017, vol. 3, pp. 2130–2143, 6 2017

    2017

  9. [17]

    Pytorch-ood: A library for out-of-distribution detection based on pytorch,

    K. Kirchheim, M. Filax, and F. Ortmeier, “Pytorch-ood: A library for out-of-distribution detection based on pytorch,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops , pp. 4351–4360, June 2022

    2022

  10. [18]

    Multi-class data description for out-of-distribution detection,

    D. Lee, S. Yu, and H. Yu, “Multi-class data description for out-of-distribution detection,” 2020

    2020

  11. [19]

    Feature selection using a multilayer perceptron,

    D. W. Ruck, S. K. Rogers, and M. Kabrisky, “Feature selection using a multilayer perceptron,” Journal of neural network computing, vol. 2, no. 2, pp. 40–48, 1990

    1990

  12. [20]

    Tabr: Tabular deep learning meets nearest neighbors in 2023,

    Y . Gorishniy, I. Rubachev, N. Kartashev, D. Shlenskii, A. Kotelnikov, and A. Babenko, “Tabr: Tabular deep learning meets nearest neighbors in 2023,” 2023

    2023

  13. [21]

    Self-normalizing neural networks,

    G. Klambauer, T. Unterthiner, A. Mayr, and S. Hochreiter, “Self-normalizing neural networks,” 31st Conference on Neu- ral Information Processing Systems , 2017

    2017

  14. [22]

    Attention is all you need,

    A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, Ł. Kaiser, and I. Polosukhin, “Attention is all you need,” Advances in neural information processing systems, vol. 30, 2017

    2017

  15. [23]

    Rest-net: Diverse activation modules and parallel subnets-based cnn for spatial image steganalysis,

    B. Li, W. Wei, A. Ferreira, and S. Tan, “Rest-net: Diverse activation modules and parallel subnets-based cnn for spatial image steganalysis,” IEEE Signal Processing Letters , vol. 25, no. 5, pp. 650–654, 2018

    2018

  16. [24]

    Dcn v2: Improved deep and cross network and practical lessons for web-scale learning to rank systems,

    R. Wang, R. Shivanna, D. Z. Cheng, S. Jain, D. Lin, L. Hong, and E. H. Chi, “Dcn v2: Improved deep and cross network and practical lessons for web-scale learning to rank systems,” The Web Conference 2021 - Proceedings of the World Wide Web Conference, WWW 2021 , pp. 1785–1797, 8 2020

    2021

  17. [25]

    Autoint: Automatic feature interaction learning via self-attentive neural networks,

    W. Song, C. Shi, Z. Xiao, Z. Duan, Y . Xu, M. Zhang, and J. Tang, “Autoint: Automatic feature interaction learning via self-attentive neural networks,” International Conference on Information and Knowledge Management, Proceedings , vol. 10, pp. 1161–1170, 10 2018

    2018

  18. [26]

    Neural oblivious decision ensembles for deep learning on tabular data,

    S. Popov, S. Morozov, and A. Babenko, “Neural oblivious decision ensembles for deep learning on tabular data,” 8th International Conference on Learning Representations, ICLR 2020, 9 2019

    2020

  19. [27]

    Tabnet: Attentive interpretable tabular learning,

    S. Arık and T. Pfister, “Tabnet: Attentive interpretable tabular learning,” 35th AAAI Conference on Artificial Intelligence, AAAI 2021, vol. 8A, pp. 6679–6687, 8 2019

    2021

  20. [28]

    Gradient boosting neural networks: Grownet,

    S. Badirli, X. Liu, Z. Xing, A. Bhowmik, K. Doan, and S. S. Keerthi, “Gradient boosting neural networks: Grownet,” 2 2020

    2020

  21. [29]

    SCARF: SELF- SUPERVISED CONTRASTIVE LEARNING USING RAN- DOM FEATURE CORRUPTION,

    D. Bahri, H. Jiang, Y . Tay, and D. Metzler, “SCARF: SELF- SUPERVISED CONTRASTIVE LEARNING USING RAN- DOM FEATURE CORRUPTION,” in ICLR 2022 - 10th In- ternational Conference on Learning Representations , 2022

    2022

  22. [30]

    Becker and R

    B. Becker and R. Kohavi, “Adult.” UCI Machine Learning Repository, 1996. DOI: https://doi.org/10.24432/C5XW20

    1996 doi

  23. [31]

    Analysis of the automl challenge series 2015-2018,

    I. Guyon, L. Sun-Hosoya, M. Boull ´e, H. J. Escalante, S. Es- calera, Z. Liu, D. Jajetic, B. Ray, M. Saeed, M. Sebag, A. Statnikov, W.-W. Tu, and E. Viegas, “Analysis of the automl challenge series 2015-2018,” in AutoML, Challenges in Machine Learning, Springer, 2019

    2015

  24. [32]

    Searching for exotic particles in high-energy physics with deep learning,

    P. Baldi, P. Sadowski, and D. Whiteson, “Searching for exotic particles in high-energy physics with deep learning,” Nature Communications, vol. 5, p. 4308, 2014

    2014

  25. [33]

    The amsterdam library of object images,

    J.-M. Geusebroek, G. J. Burghouts, and A. W. M. Smeulders, “The amsterdam library of object images,” International Jour- nal of Computer Vision , vol. 61, no. 1, pp. 103–112, 2005

    2005

  26. [34]

    Comparative accuracies of ar- tificial neural networks and discriminant analysis in predicting forest cover types from cartographic variables,

    J. A. Blackard and D. J. Dean, “Comparative accuracies of ar- tificial neural networks and discriminant analysis in predicting forest cover types from cartographic variables,”Computers and Electronics in Agriculture, vol. 24, no. 3, pp. 131–151, 2000

    2000

  27. [35]

    Sparse spatial autoregressions,

    R. K. Pace and R. Barry, “Sparse spatial autoregressions,” Statistics & Probability Letters , vol. 33, no. 3, pp. 291–297, 1997

    1997

  28. [36]

    The million song dataset,

    T. Bertin-Mahieux, D. P. Ellis, B. Whitman, and P. Lamere, “The million song dataset,” in Proceedings of the 12th Inter- national Conference on Music Information Retrieval (ISMIR 2011), (Miami, Florida, USA), pp. 591–596, October 2011

    2011

  29. [37]

    Yahoo! learning to rank challenge overview,

    O. Chapelle and Y . Chang, “Yahoo! learning to rank challenge overview,” in Proceedings of the Learning to Rank Challenge , vol. 14 of Proceedings of Machine Learning Research , pp. 1– 24, PMLR, 2011

    2011

  30. [38]

    Introducing LETOR 4.0 datasets,

    T. Qin and T.-Y . Liu, “Introducing LETOR 4.0 datasets,”arXiv preprint arXiv:1306.2597, 2013

    2013 arXiv

  31. [39]

    Tabular contrastive learning (tcl)

    A. Ginanjar, “Tabular contrastive learning (tcl).” [Online]. Available from: https://github.com/mambo06/TCL, July 12 2024

    2024

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.