Pith. sign in

REVIEW 4 major objections 6 minor 18 references

Granular Ball Twin Support Vector Machine with Universum Data

T0 review · 4 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read A twin support vector machine that represents data as granular balls and incorporates Universum samples outperforms existing TSVM variants on benchmark datasets in both accuracy and training time.

desk verdict The paper is a careful but ultimately flawed assembly of known SVM ingredients: the granular-ball radius is used as a constant offset rather than a norm-weighted margin, so the central robustness claim does not follow from the math. read the letter →

arxiv 2412.03375 v1 pith:HPB7PWBO submitted 2024-12-04 cs.LG

M. A. Ganaie , Vrushank Ahire This is my paper · ORCID
classification cs.LG
keywords TwinSupportVectorMachineGranularBallComputingUniversumdatarepresentationClassificationNoiserobustnessRBFkernel
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

This paper claims that representing training instances as granular balls—clusters with a center and a radius—and adding Universum samples (unlabeled points outside the target classes) to a twin support vector machine improves both accuracy and training time. The proposed GBU-TSVM replaces point-wise constraints with ball-wise constraints in two non-parallel hyperplane problems. On ten benchmark datasets it reports the highest mean accuracy in eight cases, and for datasets near 1000 samples it reports training roughly 16 times faster than U-TSVM and 58 times faster than TSVM. If correct, this would make ball-based representations a practical upgrade for twin SVMs on noisy or larger datasets.

What carries the argument

The central object is the granular ball, a hyper-ball with centroid $c$ and radius $r$, which replaces point instances in the twin SVM constraints. Its radius enters the constraints as a constant offset—for example, $-(B\omega_+ + e_- b_+) + \xi \ge e_- - R_-$—rather than scaled by $\|\omega\|$ as in the original GB-SVM. Universum samples enter through an $\epsilon$-insensitive tube with hinge losses $\psi$ and $\psi^*$, pulling the two non-parallel hyperplanes away from the unlabeled data. The dual quadratic programs for the two hyperplanes, and the minimum-distance rule, are the mechanism that turns this representation into a classifier.

What would settle it

Compute the original GB-SVM's weighted-radius constraint on the same granular balls used in the experiments and compare test accuracy against GBU-TSVM's unweighted version; if the weighted version matches or beats GBU-TSVM while the unweighted version underperforms, the simplification is the source of the gain. Alternatively, add an outlier far from a ball center and check whether the decision boundary changes as the outlier distance grows while the ball radius stays fixed: the unweighted constraint predicts no change, the weighted constraint predicts a shift.

Watch

Extended reading notes

Core claim

The central claim is that the twin SVM's two quadratic programs can be written with granular balls as inputs, where each ball is a centroid and a radius, and with Universum data placed in an insensitive loss tube, yielding a model that is more robust to noise and cheaper to train than point-based TSVM variants. The paper derives linear and RBF-kernel versions of the primal and dual problems, and reports experiments where GBU-TSVM achieves the best average accuracy on 8 of 10 benchmark datasets and substantial speedups. The authors position GBU-TSVM as a more general formulation from which U-TSVM, Pin-GTSVM, and other TSVM models can be derived by modifications.

Load-bearing premise

The model assumes a ball's radius can be subtracted as a fixed amount from the margin requirement, rather than scaled by the length of the separating plane's normal vector; if this simplification is wrong, the claimed noise resistance and accuracy gains may not follow.

Editorial extensions

If this is right

  • If the reported accuracies hold, granular-ball representation becomes a practical preprocessing step that lifts TSVM-family classifiers on noisy and imbalanced benchmarks.
  • The reported speedups—16 times versus U-TSVM and 58 times versus TSVM on datasets near 1000 samples—imply that ball reduction makes twin SVMs feasible on data sizes where point-based quadratic programs were too slow.
  • The Universum tube gives a concrete way to inject unlabeled in-domain data into pairwise-margin classifiers, potentially improving calibration on small labeled sets.
  • Since balls summarize many points by a centroid and radius, the learned decision boundary has a natural interpretation in terms of groups of points rather than individual support vectors.

Reading between the lines

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

  • The paper's decision to drop the $\|\omega\|$ scale on the radius is never derived; a direct comparison with the original weighted-radius GB-SVM on the same datasets would isolate whether the accuracy gain comes from ball representation or from the simplified constraint.
  • The nonlinear section asserts balls are generated in kernel space but does not specify how the kernel-induced centroids and radii are computed; an implementable recipe is needed to verify the claim for RBF kernels.
  • The reported speedup depends on the number of balls produced by the minimum-count and purity thresholds; reporting training time as a function of ball count would let practitioners predict when the method pays off.
  • Because Universum data is often scarce in practice, synthesizing Universum balls from the training distribution—an approach the paper tried and set aside as computationally expensive—is a natural testable extension.
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 / 6 minor

Summary. The paper proposes GBU-TSVM, a twin support vector machine variant that replaces point samples with granular-ball centroids and radii and incorporates Universum data. For the linear case, the authors state two primal QPPs in which the granular-ball radii appear as constant right-hand-side offsets, derive the corresponding Wolfe duals via KKT conditions, and repeat the construction for an RBF-kernel version. Experiments on ten UCI datasets report higher mean accuracy than U-TSVM, TSVM, and Pin-GTSVM in most cases, together with claims of large speedups and statistical significance.

Significance. If the model performed as claimed, combining granular-ball representations with Universum learning inside the TSVM framework would be a useful contribution to scalable and noise-tolerant classification. The dual derivations in the linear case are algebraically transparent, and the idea of reducing training data to ball centroids/radii is attractive. However, the central modeling step does not actually enforce robust hyper-ball separation, and the nonlinear extension mixes feature-space centers with input-space kernel matrices. The current manuscript therefore establishes only a heuristic radius-offset TSVM variant, not a robust granular-ball classifier, so the scientific significance is not yet established.

major comments (4)
  1. [Section 3.1, Eqs. (54)-(55)] The granular-ball radius enters the constraints as a constant right-hand-side offset, e.g., -(Bω+ + e- b+) + ξ ≥ e- - R-, rather than through the norm-weighted term ||ω|| r used in GB-SVM's constraint (7). For a negative-class ball with center c and radius r, requiring every point in the ball to satisfy the margin condition ω+^T x + b+ ≤ -1 is equivalent to -(ω+^T c + b+) ≥ 1 + ||ω+|| r, not ≥ 1 - r. With the paper's formulation the required margin shrinks as r grows and can even become negative for r>1, allowing the ball center to lie on the wrong side of the hyperplane. The dual in Eq. (72) is consistent with this non-robust primal, but the manuscript's claims of noise resistance and hyper-ball robustness do not follow from the stated model.
  2. [Section 3.2, Eqs. (84), (91), (94)] In the nonlinear case, centers c_j and radii r_j are defined in the feature space, with c_j = (1/k) Σ φ(z_i) in Eq. (84). However, the hyperplanes in Eqs. (91)-(92) and the constraints in Eqs. (93)-(96) use kernel matrices K(A,C^T), K(B,C^T), K(U,C^T), where C^T = [A B]^T. If C contains input-space centroids, then the feature-space centers of Eq. (84) are not actually used; if C is intended to contain feature-space centers, then K(·, C^T) is not a well-defined kernel matrix between input-space vectors. The nonlinear model therefore does not implement the stated feature-space granular-ball construction, and the derivation of the dual in Eqs. (97)-(101) inherits this inconsistency.
  3. [Section 4.3] The experimental comparison is not sufficiently supported. The text states that U-TSVM, Pin-GTSVM, and other TSVM baselines 'were derived by making necessary modifications to the GBU-TSVM model to record the accuracies.' This raises the risk that the baselines were reimplemented from the proposed model's code without validating them against the original published implementations. In addition, Table 2 reports only mean accuracies without standard deviations, the claimed speedups ('16 times faster than U-TSVM and 58 times faster than TSVM') are not accompanied by a runtime table, and the text states that hyperparameter tuning time is excluded. These issues are load-bearing for the abstract's accuracy and efficiency claims.
  4. [Section 4.2] The procedure for choosing the granular-ball parameters num and pur is described as first 'observed and recorded for each dataset' to select favorable ranges, which risks information leakage from the test split. The relationship between the 50% training / 30% universum / 20% test split and the 5-fold cross-validation for hyperparameter selection is also not explained clearly. A precise protocol that separates tuning data from evaluation data is needed before the reported performance can be considered credible.
minor comments (6)
  1. [Section 4.3] There is a typo in Section 4.3: 'GBY-TSVM' should be 'GBU-TSVM.'
  2. [Tables 1-2 and Section 4.3] The dataset sizes are inconsistent: Table 1 lists Oocytes-Trisopterus-Nucleus-2f as 912 samples and Oocytes-Merluccius-Nucleus-4d as 1022 samples, but Section 4.3 refers to 'Oocytes-Trisopterus dataset (1022 samples)' and Table 2 lists only 'Oocytes-Trisopterus' with accuracy 83.06. Please reconcile the dataset naming and sizes.
  3. [Section 4.3.5] The parameter sensitivity analysis reports ϵ = 32.0 as the top value, but the experimental setup in Section 4.2 restricts ϵ to (0,1) and to the set {0, 0.2, 0.4, 0.6, 0.8, 1}. This inconsistency needs clarification.
  4. [Eq. (111)] The Wilcoxon signed-rank statistic is written as W = Σ sign(d_i) R_i, which is not the standard paired signed-rank statistic used in the reported p-values. Please state the exact test statistic and software used.
  5. [Section 4.2] The hyperparameter grid is described as {2^{2i} | i = -4, ..., 4} in Section 4.2 but as {2^k | k = -8, ..., 8} in Section 4.3.5; please make the grid definitions consistent.
  6. [Section 3.1, Eqs. (42)-(43), (55)] The definition of Universum radii R_u is introduced but its role is only briefly used in the constraints; please clarify how the radii of Universum granular balls are computed and how they affect the loss in practice.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the GBU-TSVM dual problems are derived from the stated primals by standard KKT algebra, and the reported comparisons are empirical rather than derived.

full rationale

The paper's optimization derivation is self-contained. Starting from the explicitly stated primal QPPs (54)-(55) and (56)-(57), the Lagrangian in (58) is formed, the KKT conditions in (62)-(65) yield the stationarity equations, and the Wolfe duals in (72) and (74) follow by substitution. The radius terms R_-, R_+, and R_u enter as constant right-hand-side offsets; this is a modeling-validity concern relative to the norm-weighted GB-SVM constraint in Eq. (7), but it is not circular because the dual is algebraically consistent with the paper's stated primal. The kernel extension in (93)-(96) and its duals in (97)-(102) repeat the same derivation, and no equation is defined in terms of the result it is supposed to predict. The experimental section reports measured accuracies and timings; the statements that U-TSVM and Pin-GTSVM were reimplemented by modifying the proposed model and that num/pur ranges were observed before tuning are experimental fairness and reproducibility concerns, not circular reductions of the central claim. There is no load-bearing self-citation: the GB-SVM and granular-ball references are external prior work, and no fitted parameter or uniqueness theorem is renamed as a prediction. Therefore the derivation chain is not circular.

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

The central claim depends on two fitted hyperparameter families, regularization and Universum weights plus ball-generation thresholds, an unproven assumption that Universum data help, and an ad hoc treatment of radii as constant offsets rather than norm-weighted margin terms. No new physical or mathematical entities are introduced.

free parameters (5)
  • c1, c2, cu = selected per dataset by 5-fold CV; sensitivity analysis top values c1=8.0, c2=4.0, cu=0.25
    Regularization and Universum trade-off parameters in QPPs (54) and (56).
  • epsilon = selected from {0, 0.2, 0.4, 0.6, 0.8, 1}; sensitivity analysis reports 32.0, contradicting the stated range
    Controls the Universum insensitive tube in constraints (55) and (57).
  • num (minimum points per granular ball) = not reported per dataset; chosen after observing ball generation
    Threshold for ball generation used in all experiments.
  • pur (purity threshold) = not reported per dataset; chosen after observing ball generation
    Threshold for ball generation quality used in all experiments.
  • RBF kernel width sigma = not reported in the experimental setup
    Width of the Gaussian kernel used in the nonlinear formulation, Eq. (82).
assumptions (4)
  • standard math KKT conditions and Wolfe duality produce the dual from the primal QPP.
    Used in Section 3.1, Eq. (59)-(71), to derive the dual problems; standard for convex quadratic programming.
  • domain assumption Universum samples carry useful prior information for classification.
    Borrowed from Weston et al. [6] and Qi et al. [7]; not proven for the specific UCI datasets used here.
  • ad hoc to paper Granular-ball radius can enter the constraints as a constant RHS offset without a ||omega|| weighting term.
    Used in Eq. (54)-(57) and Eq. (93)-(96); diverges from GB-SVM Eq. (7) and is not derived.
  • ad hoc to paper Feature-space granular-ball centers can be handled with ordinary kernel matrices between input-space vectors.
    Section 3.2 defines centers as averages of phi(z_i) in feature space, Eq. (84), but then treats them as vectors in R^n in Eq. (87)-(94).

how reviews work

0 comments
Cite this review

Pith. "Pith review of Granular Ball Twin Support Vector Machine with Universum Data." pith.science (2026). https://pith.science/paper/HPB7PWBO

@misc{pith2026241203375,
  author       = {Pith},
  title        = {Pith review of: Granular Ball Twin Support Vector Machine with Universum Data},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/HPB7PWBO}},
  note         = {Machine review of arXiv:2412.03375}
}
read the original abstract

Classification with support vector machines (SVM) often suffers from limited performance when relying solely on labeled data from target classes and is sensitive to noise and outliers. Incorporating prior knowledge from Universum data and more robust data representations can enhance accuracy and efficiency. Motivated by these findings, we propose a novel Granular Ball Twin Support Vector Machine with Universum Data (GBU-TSVM) that extends the TSVM framework to leverage both Universum samples and granular ball computing during model training. Unlike existing TSVM methods, the proposed GBU-TSVM represents data instances as hyper-balls rather than points in the feature space. This innovative approach improves the model's robustness and efficiency, particularly in handling noisy and large datasets. By grouping data points into granular balls, the model achieves superior computational efficiency, increased noise resistance, and enhanced interpretability. Additionally, the inclusion of Universum data, which consists of samples that are not strictly from the target classes, further refines the classification boundaries. This integration enriches the model with contextual information, refining classification boundaries and boosting overall accuracy. Experimental results on UCI benchmark datasets demonstrate that the GBU-TSVM outperforms existing TSVM models in both accuracy and computational efficiency. These findings highlight the potential of the GBU-TSVM model in setting a new standard in data representation and classification.

Figures

Figures reproduced from arXiv: 2412.03375 by the authors.

Figure 1
Figure 1. Global precedence in human cognition Granular-ball Support Vector Machine (GB-SVM) (Xia et al., 2019)[2], an innova￾3 [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Visualizing Universum Data and Classes the development of the Universum TSVM (U-TSVM) (Qi et al., (2012) [7]). Unlike traditional U-SVM, U-TSVM utilizes two hinge loss functions to place Universum data in a non-parallel insensitive loss tube, allowing for more flexible exploitation of prior knowledge. This approach demonstrated superior performance compared to traditional SVMs in various tasks. Recognizing the poten… view at source ↗
Figure 3
Figure 3. Hinge Loss ψ 18 [PITH_FULL_IMAGE:figures/full_fig_p018_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Hinge Loss ψ ∗ The Lagrangian associated with equations (54)–(55) can be expressed as: L(Ω) = 1 2 ∥Aω+ + e+b+∥ 2 + c1e T −ξ + cue T uψ − α T (−(Bω+ + e−b+) + ξ − e− + R−) − β T ξ − µ T ((Uω+ + eub+) + ψ + (1 − ϵ)eu + Ru) − γ Tψ, (58) where Ω = {ω+, b+, ξ, ψ, α, β, µ, γ…
Figure 5
Figure 5. Figure 5: GBU-TSVM Geometric Interpretation challenging because the radius cannot be directly mapped. Mapping original space balls to kernel space changes the data distribution, making them inaccurate. The best solution is to generate granular-balls directly in the kernel space.…
Figure 6
Figure 6. Figure 6: Model Performance Across Various Dataset Sizes [PITH_FULL_IMAGE:figures/full_fig_p028_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

18 extracted references · 16 canonical work pages

  1. [1]

    Granular ball computing: an ef- ficient, robust, and interpretable adaptive multi-granularity representation and computation method

    Shuyin Xia, Guoyin Wang, and Xinbo Gao. Granular ball computing: an ef- ficient, robust, and interpretable adaptive multi-granularity representation and computation method. arXiv preprint arXiv:2304.11171, 2023

    arXiv 2023

  2. [2]

    Granular ball computing classifiers for efficient, scalable and robust learning

    Shuyin Xia, Yunsheng Liu, Xin Ding, Guoyin Wang, Hong Yu, and Yuoguo Luo. Granular ball computing classifiers for efficient, scalable and robust learning. Information Sciences, 483:136–152, 2019

    work page 2019

  3. [3]

    Twin support vector machines for pattern classification

    Jayadeva, Reshma Khemchandani, Suresh Chandra, et al. Twin support vector machines for pattern classification. IEEE Transactions on pattern analysis and machine intelligence, 29(5):905–910, 2007

    work page 2007

  4. [4]

    Application of smoothing technique on twin support vector machines

    M Arun Kumar and Madan Gopal. Application of smoothing technique on twin support vector machines. Pattern Recognition Letters, 29(13):1842–1848, 2008

    work page 2008

  5. [5]

    A coordinate descent margin based-twin support vector machine for classification

    Yuan-Hai Shao and Nai-Yang Deng. A coordinate descent margin based-twin support vector machine for classification. Neural networks, 25:114–121, 2012

    work page 2012

  6. [6]

    Inference with the universum

    Jason Weston, Ronan Collobert, Fabian Sinz, L´ eon Bottou, and Vladimir Vap- nik. Inference with the universum. In Proceedings of the 23rd international conference on Machine learning, pages 1009–1016, 2006

    work page 2006

  7. [7]

    Twin support vector machine with universum data

    Zhiquan Qi, Yingjie Tian, and Yong Shi. Twin support vector machine with universum data. Neural Networks, 36:112–119, 2012. 33

    work page 2012

  8. [8]

    Least squares twin support vector machine with universum data for classification

    Yitian Xu, Mei Chen, and Guohui Li. Least squares twin support vector machine with universum data for classification. International Journal of Systems Science, 47(15):3637–3645, 2016

    work page 2016

Show all 18 references
  1. [9]

    ν-twin support vector machine with universum data for classification

    Yitian Xu, Mei Chen, Zhiji Yang, and Guohui Li. ν-twin support vector machine with universum data for classification. Applied Intelligence, 44(4):956–968, 2016

    2016

  2. [10]

    A reduced universum twin support vector machine for class imbalance learning

    Bharat Richhariya and Muhammad Tanveer. A reduced universum twin support vector machine for class imbalance learning. Pattern Recognition, 102:107150, 2020

    2020

  3. [11]

    Uni- versum parametric-margin ν-support vector machine for classification using the difference of convex functions algorithm

    Hossein Moosaei, Fatemeh Bazikar, Saeed Ketabchi, and Milan Hlad ´ ık. Uni- versum parametric-margin ν-support vector machine for classification using the difference of convex functions algorithm. Applied Intelligence, 52(3):2634–2654, 2022

    2022

  4. [12]

    Inverse free reduced universum twin support vector machine for imbalanced data clas- sification

    Hossein Moosaei, MA Ganaie, Milan Hlad ´ ık, and Muhammad Tanveer. Inverse free reduced universum twin support vector machine for imbalanced data clas- sification. Neural Networks, 157:125–135, 2023

    2023

  5. [13]

    A new multi-task learning method with universum data

    Yanshan Xiao, Jing Wen, and Bo Liu. A new multi-task learning method with universum data. Applied Intelligence, 51:3421–3434, 2021

    2021

  6. [14]

    A new transductive learning method with universum data

    Yanshan Xiao, Junyao Feng, and Bo Liu. A new transductive learning method with universum data. Applied Intelligence, 51(8):5571–5583, 2021

    2021

  7. [15]

    Universum based lagrangian twin bounded support vector machine to classify eeg signals

    Bikram Kumar and Deepak Gupta. Universum based lagrangian twin bounded support vector machine to classify eeg signals. Computer methods and programs in biomedicine, 208:106244, 2021

    2021

  8. [16]

    Adaptive robust adaboost-based twin support vector machine with universum data

    Bo Liu, Ruiguang Huang, Yanshan Xiao, Junrui Liu, Kai Wang, Liangjiao Li, and Qihang Chen. Adaptive robust adaboost-based twin support vector machine with universum data. Information Sciences, 609:1334–1352, 2022

    2022

  9. [17]

    An improved parametric-margin universum tsvm

    Yanmeng Li and Huaijiang Sun. An improved parametric-margin universum tsvm. Neural Computing and Applications, 34(16):13987–14001, 2022

    2022

  10. [18]

    Gbsvm: granular-ball support vector machine

    Shuyin Xia, Guoyin Wang, Xinbo Gao, and Xiaoli Peng. Gbsvm: granular-ball support vector machine. arXiv preprint arXiv:2210.03120, 2022. 34

    2022 arXiv

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.