- Rating
- 3(+1)
- Confidence
- 3
- Informativeness
- 2
- Presentation
- 3
- Soundness
- 2
Read review
Paper summary
This paper proposes Tabular-to-Graph Diffusion (TGD) for minority-class oversampling. It represents local minority neighborhoods as feature graphs and combines graph convolutions with localized diffusion sampling. Experiments on eight datasets show improved balanced accuracy and competitive performance against traditional oversampling and tabular generative methods.
Critical strengths
-
I liked the idea of combining local neighborhood information with feature-level graph representations. This is a reasonable way to preserve feature dependencies when minority samples are limited.
-
The experimental evaluation is fairly comprehensive. The authors compare against several strong baselines and include ablation studies, distribution visualizations, and runtime comparisons.
Critical weaknesses
-
I am not fully convinced that the improvement comes mainly from the proposed graph representation. The method also uses localized warm-start sampling and post-processing, which can independently improve generation quality. Although the paper includes ablations, a matched vector-diffusion baseline using the same sampling strategy would make the contribution clearer.
-
The improvement over SMOTE seems relatively limited given the additional model complexity. Although TGD achieves better average ranks, the Nemenyi analysis does not establish a significant difference from SMOTE. The statistical analysis also treats multiple correlated metrics from the same datasets as separate tasks, which makes the reported significance harder to interpret.
Questions for authors
-
How much of the improvement comes from the graph representation rather than localized sampling? Could a standard vector diffusion model with the same warm-start strategy achieve similar results?
-
Since TGD does not significantly outperform SMOTE under the reported Nemenyi test, what is the main practical advantage that justifies the additional computational cost?
-
Would the statistical conclusions change if significance were evaluated across the eight independent datasets rather than treating all 24 dataset-metric combinations as separate tasks?
4 comments
### Response to W1 & Q1: Comparing Full TGD against 1D Vector Diffusion under Matched Sampling
Thank you for raising this point regarding performance attribution. To investigate whether a standard vector diffusion model with localized sampling could achieve comparable performance (Q1), we implemented the exact matched 1D vector baseline in Appendix H.3 (Tables 7–8).
Our empirical results show that it falls short: even with identical localized warm-start perturbation and post-processing, standard vector diffusion degrades, with its overall mean rank dropping from 1.645 to 6.479 across all eight benchmark datasets.
The strictly matched comparison from the extended ablation suite (Appendix H.3, Tables 7–8) is summarized below:
| Method / Variant | Feature Architecture | Overall Mean Rank (Tables 7–8) | Phoneme () Bal-Acc (Rank) | Churn () Bal-Acc (Rank) | Ol8 () Bal-Acc / AUC / F1 (Rank) |
|---|---|---|---|---|---|
| Full TGD (Ours) | Transposed Patch + GCN | 1.645 | 0.843 (1.000) | 0.830 (1.000) | 0.741 / 0.817 / 0.388 (2.333) |
| Matched 1D Vector Diff. | Flat Vector + MLP | 6.479 | 0.833 (8.667) | 0.814 (7.000) | 0.716 / 0.792 / 0.371 (7.667) |
- Controlled Experimental Alignment (Appendix H.3): To isolate architectural contributions from localized sampling and post-processing, the Matched 1D Vector baseline strictly shares:
- Equivalent Capacity: A standard residual MLP backbone with matched hidden dimensions () operating on flattened 1D records ;
- Identical Sampling: The exact same localized warm-start perturbation (, ) and deterministic reverse PF-ODE integration ();
- Identical Post-Processing: Empirical moment matching, coordinate interval clamping, and categorical Argmax discretization.
- Disentangling the Gain Hierarchy (Connecting Table 6 and Tables 7–8): Analyzing these ablation suites reveals a clear, complementary two-level contribution hierarchy across representation and message passing:
- Level 1 (Local Coordinate Smoothing via Transposed Patch): In the representation ablation suite (9 variants, Tables 7–8), transitioning from flat 1D vectors to structured transposed patches accounts for the primary performance gain, improving the overall mean rank from 6.479 (Matched 1D) to 1.645 (Full TGD). Organizing ordered local neighbors along the coordinate axis allows linear projection layers () to smooth variations across neighboring minority records, which isolated 1D MLPs cannot capture.
- Level 2 (Column Covariance Regularization via Graph Convolutions): In the architectural ablation suite (7 variants, Table 6), incorporating correlation-guided GCN layers over the transposed patch provides the remaining critical gains, advancing overall rank from 4.667 ('w/o Graph') to 2.291 (Full TGD). As analyzed in Section 5.3, while neighbor-axis projection alone maintains local support, graph message passing explicitly penalizes discordant column updates (minimizing Dirichlet energy), preventing the severe high-dimensional covariance degradation observed in flat diffusion (e.g., raising Balanced Accuracy on Ol8 from 0.716 to 0.738 in Table 6, versus 0.716 for Matched 1D and 0.741 for Full TGD in Table 8).
These comparisons verify that localized sampling provides a reliable neighborhood anchor, but the transposed graph architecture is essential for preserving multi-feature dependencies and achieving top-tier classification fidelity. Complete quantitative metrics are reported in Appendix H.1 (Table 6) and Appendix H.3 (Tables 7–8).
### Response to W2 & Q2: Practical Advantages over SMOTE: Decision Boundaries and Feature Correlations
Thank you for this perceptive assessment regarding the comparison with SMOTE and the statistical evaluation protocol.
We fully share your respect for SMOTE. In tabular learning, classical heuristics like SMOTE remain remarkably competitive, mirroring the ongoing dynamic between tree-based ensembles (GBDTs) and deep tabular architectures. While neural diffusion inevitably introduces higher computational overhead than linear interpolation, this modest cost delivers essential geometric and structural capabilities that heuristic methods cannot provide.
The practical advantages of TGD over SMOTE, alongside our updated statistical evaluation, are summarized below:
| Evaluation Dimension | SMOTE (Heuristic Baseline) | TGD (Transposed Graph Diffusion) | Practical Advantage of TGD |
|---|---|---|---|
| Statistical Significance (, Bal-Acc) | Mean Rank: 2.500 (24-Task: 3.167) | Mean Rank: 1.125 (24-Task: 1.875) | Reject (, Wilcoxon) |
| Boundary Fidelity | Linear straight-line segments | Bounded reverse ODE trajectories | Eliminates crossing into majority clusters |
| Feature Covariance | Unmodeled (independent axes) | Graph convolutions () | Preserves multi-column dependencies |
| Categorical Validity | Fractional interpolation values | + Categorical Argmax | Enforces strict one-hot mutual exclusivity |
| Computational Footprint | < 1 second (CPU) | 14.6s–92.5s (, Laptop GPU) | Real-time feasible without pre-training |
-
Statistical Significance on Independent Datasets (): The global Nemenyi test is conservative when controlling family-wise error across all 45 algorithm pairs (). When evaluating Balanced Accuracy strictly across the independent dataset blocks, TGD achieves an average rank of 1.125 (ranking 1st on 7 datasets and 2nd on SD) compared to 2.500 for SMOTE, yielding () against SMOTE under a two-sided Wilcoxon signed-rank test (Appendix C.5). Furthermore, evaluating net improvements across all 24 evaluation tasks (Appendix F.3, Table 4) confirms that TGD maintains the leading overall rank of 1.875 with an average gain of over raw training, compared to 3.167 and for SMOTE.
-
Elimination of Decision Boundary Violations (Boundary Fidelity): SMOTE connects neighboring minority instances via unconstrained straight lines. On non-convex or multi-modal distributions, these segments cut across intermediate majority-class valleys, introducing severe label noise (Figure 4 and Appendix Figure 8). On datasets like SD and BCCP, this causes synthetic points to scatter into majority territory, dropping SMOTE's F1 rank to 8th on SD. In contrast, TGD perturbs local patches at and integrates inward, confining synthetic samples strictly within authentic minority clusters.
-
Preserving Feature Correlations and Discrete States: Linear interpolation treats attributes independently, failing to preserve column covariance and producing fractional floats on one-hot indicators. TGD preserves pairwise feature dependencies via correlation-guided GCN layers and strictly enforces discrete validity via categorical masking and post-hoc Argmax discretization, scaling reliably up to 299 features (Musk and Scene, Appendix H.4, Table 9).
-
Practical Runtime: Unlike global tabular diffusion models that incur quadratic attention overhead across columns, TGD operates on sparse local graphs () without pre-training. Computational cost is governed primarily by dataset sample size and the target synthesis volume rather than feature dimensionality . Consequently, TGD scales efficiently to high-dimensional tables, requiring only 14.62 seconds on Scene () and 92.52 seconds on Musk () on a standard consumer RTX 3060 Laptop GPU (Appendix H.4, Table 9).
### Response to Q3: Statistical Significance Evaluated Across N=8 Independent Datasets
Thank you for pointing out the issue of metric correlation. In our original formulation, evaluating across 24 dataset-metric tasks was intended to provide a comprehensive, multi-faceted benchmark overview across complementary performance criteria (macro balance, global ranking, and minority precision/recall trade-offs). Nevertheless, we fully agree with your perspective that from a rigorous mathematical statistics standpoint, eliminating potential inter-metric co-dependencies is essential.
Accordingly, in Appendix C.5, we have re-executed the statistical testing protocol by treating the eight distinct datasets strictly as independent evaluation blocks (), and find that our statistical conclusions remain consistent.
The re-evaluated non-parametric statistical tests across the independent datasets are summarized below:
| Statistical Test Protocol | Evaluation Scope | Test Statistic | p-value | Decision / Verdict |
|---|---|---|---|---|
| Friedman Test | Balanced Accuracy alone | Reject (Difference among 10 methods) | ||
| Friedman Test | AUC-ROC alone | Reject (Difference among 10 methods) | ||
| Friedman Test | Dataset-Level Mean Rank | Reject (Difference among 10 methods) | ||
| Wilcoxon Signed-Rank | TGD vs. SMOTE (Bal-Acc) | Reject at () | ||
| Wilcoxon Signed-Rank | TGD vs. Other 8 Baselines | Reject at () |
-
Multi-Classifier ANOVA (Friedman): Whether testing Balanced Accuracy alone, AUC alone, or collapsing all metrics into a single unweighted dataset-level rank, the Friedman test rejects the null hypothesis of equivalent performance across the 10 algorithms ().
-
Pairwise Comparisons against Baselines (Wilcoxon): Treating datasets as independent blocks, the two-sided Wilcoxon signed-rank test on Balanced Accuracy yields with ( under one-sided testing) against SMOTE, rejecting the null hypothesis at . Against each of the remaining eight baselines (Original, ADASYN, CTGAN, TVAE, TabDDPM, GOGGLE, TabSyn, and TabDiff), the test yields with (two-sided).
Therefore, eliminating inter-metric correlation and restricting statistical tests strictly to independent dataset blocks leads to the same conclusions ( across all pairwise comparisons). Full details are documented in Appendix C.5.
### General Response and Summary of Revisions
Thank you for your constructive review and insightful questions on attribution, practical utility, and statistical rigor. Following your suggestions, we have revised the manuscript to include:
(1) a strictly aligned
Matched 1D Vector Diffusionablation across all eight benchmark datasets (sharing the exact same localized warm-start perturbation twarm=30, deterministic PF-ODE integration, and post-processing) to isolate the contribution of the transposed feature-graph representation (Appendix H.3, Figure 11, Tables 7–8);(2) block-independent non-parametric statistical tests evaluated strictly across the N=8 independent datasets to eliminate inter-metric correlation, yielding p=0.0156 against SMOTE and p=0.0078 against the remaining baselines under two-sided Wilcoxon signed-rank tests (Section 5.1; Appendix C.5);
(3) relative net gain analyses over unaugmented training (Δ in Appendix F.3, Table 4) alongside high-dimensional scalability benchmarks up to D=299 (Appendix H.4, Table 9), further clarifying TGD's practical advantages over linear interpolation in boundary adherence, multi-feature covariance preservation, and categorical validity. Below are our point-by-point responses.