- Rating
- 3
- Confidence
- 2
- Informativeness
- 2
- Presentation
- 3
- Soundness
- 3
Read review
Paper summary
The paper introduces degree-potential calibration (DPC), which adjusts sparse GCN propagation as , where and is a centered degree-discrepancy coordinate. An optional signed spectral calibration (SSC) applies , with one bounded learned scalar per layer, to smooth or sharpen the resulting state. The analysis establishes separation of shared and local corrections and constructs labelings with identical spectral energies and initial margin distributions but opposite locally beneficial calibration signs. Across 16 benchmarks, DPC matches or exceeds the highest baseline mean on all eight homophilous datasets, while DPC-GCN-S leads on four of five heterophilous datasets. Ablations examine the coordinate construction, sign, parameter granularity, and resource costs.
Critical strengths
-
A precise, economical construction with substantive analysis. The centered coordinate makes the local correction orthogonal to the shared mixing direction in the Frobenius geometry, giving a clean interpretation of the additional parameter. The spectral response of SSC is explicit. More unusually, Proposition 3 controls both spectral energy and the initial margin distribution while reversing the beneficial risk derivative. Its explicit construction is internally consistent under an independent numerical check at . This is a useful caution against choosing a smoothing/sharpening direction solely from aggregate spectral statistics.
-
Broad evaluation with relevant design controls. The experiments include tuned local GNNs and modern graph Transformers across graph types and scales. Appendices B.2–B.3 compare centered, uncentered, shuffled, sign-constrained, feature-wise, node-wise, and fixed-propagation variants. The paper also reports efficiency at the selected operating points and the accuracy deficit on ogbn-products. These controls make the evidence more informative than an aggregate leaderboard alone.
Critical weaknesses
-
The empirical necessity of the degree-dependent local correction is limited. Proposition 1 establishes geometric non-redundancy, which does not itself establish improved prediction. In Table 5, removing DPC while retaining SSC gives exactly the same displayed mean on Roman-Empire and Questions, and costs only and points on Minesweeper and Tolokers. Table 6 also shows that graph-wide or shuffled-centered controls are competitive on several tasks. Thus the strongest empirical case for the complete construction is concentrated on a subset of datasets. This limits how much the results establish about the distinctive degree-dependent component, even though the compact combined method remains useful.
-
Small leaderboard gains need stronger control of model-selection uncertainty. Several improvements over the strongest baselines are fractions of a percentage point. Appendix C.3 selects degree-coefficient initializations and SSC learning-rate multipliers, while Appendix C.4 uses different candidate/configuration policies for different baselines. The paper is transparent about these choices, but does not give comparable candidate counts or paired uncertainty for the decisive accuracy differences. The resource comparisons at separately selected widths/depths are useful operating-point comparisons, rather than an isolated measurement of the calibration operator's overhead. These issues limit the strength of a general superiority claim; they do not invalidate the reported ablations or the spectral derivations.
Questions for authors
-
Evidence request: the value of the local degree coordinate. Using the same run/split pairs, what are the accuracy differences and uncertainty intervals between full DPC, graph-wide mixing, and shuffled-centered coordinates in Table 6, particularly on Computer, WikiCS, and Minesweeper? This would clarify whether the reported advantage of the topology-specific local correction is consistent across runs.
-
Clarification: the contribution of DPC beyond SSC. In Table 5, retaining SSC while removing DPC gives the same displayed mean on Roman-Empire and Questions and only slightly lower means on Minesweeper and Tolokers. Which empirical results most strongly support the incremental benefit of DPC once SSC is present? Please distinguish evidence for the degree-calibration component from evidence for the combined architecture.
-
Clarification: comparable model-selection effort. What are the effective numbers of configurations evaluated for DPC and the strongest tuned GNN baselines, including coefficient initializations and SSC learning-rate multipliers? Since several leaderboard margins are small, please explain whether the conclusions remain supported after accounting for differences in model-selection effort.
-
Evidence request: efficiency attributable to calibration. The existing resource comparisons use each method's selected configuration. At matched hidden width, depth, and training protocol, what additional time and memory do DPC and SSC introduce relative to the underlying GCN? Please distinguish this operator-level overhead from the useful accuracy–resource trade-offs obtained with separately selected architectures.
3 comments
2. DPC beyond SSC
To distinguish DPC's contribution from that of the combined architecture, we compare four configurations with the same backbone: Backbone, DPC-GCN (Backbone + DPC), Backbone + SSC, and DPC-GCN-S (Backbone + DPC + SSC). Backbone uses standard GCN propagation. Normalization, residual connections, width, depth, dropout and training budget are unchanged; only DPC and SSC are enabled or disabled. Results are five-run means ± standard deviations on a percentage scale: ROC-AUC for Minesweeper, Questions and Tolokers, and accuracy otherwise.
Homophilous datasets
| Model / control | Cora | CiteSeer | PubMed | Computer | Photo | CS | Physics | WikiCS |
|---|---|---|---|---|---|---|---|---|
| Backbone | 84.46 ± 0.60 | 72.40 ± 0.47 | 81.30 ± 0.57 | 93.74 ± 0.16 | 95.88 ± 0.21 | 96.16 ± 0.14 | 97.47 ± 0.10 | 80.17 ± 0.57 |
| DPC-GCN | 85.22 ± 0.90 | 73.28 ± 0.60 | 81.36 ± 0.51 | 94.20 ± 0.10 | 96.54 ± 0.12 | 96.26 ± 0.10 | 97.47 ± 0.05 | 80.90 ± 0.28 |
| Backbone + SSC | 84.98 ± 0.44 | 72.76 ± 0.62 | 81.58 ± 0.33 | 93.86 ± 0.19 | 95.59 ± 0.21 | 96.15 ± 0.10 | 97.43 ± 0.04 | 80.29 ± 0.23 |
| DPC-GCN-S | 85.26 ± 0.64 | 73.08 ± 0.54 | 81.42 ± 0.39 | 94.21 ± 0.22 | 96.51 ± 0.13 | 96.25 ± 0.11 | 97.46 ± 0.05 | 80.83 ± 0.25 |
Heterophilous datasets
| Model / control | Ratings | Minesweeper | Roman-Empire | Questions | Tolokers |
|---|---|---|---|---|---|
| Backbone | 54.21 ± 0.51 | 97.60 ± 0.49 | 91.38 ± 0.35 | 79.06 ± 0.49 | 86.37 ± 0.54 |
| DPC-GCN | 55.42 ± 0.38 | 97.73 ± 0.38 | 91.58 ± 0.22 | 79.06 ± 0.54 | 86.41 ± 0.56 |
| Backbone + SSC | 54.35 ± 0.16 | 98.58 ± 0.32 | 91.97 ± 0.16 | 79.89 ± 0.63 | 86.52 ± 0.57 |
| DPC-GCN-S | 55.55 ± 0.39 | 98.73 ± 0.19 | 91.97 ± 0.22 | 79.89 ± 0.63 | 86.62 ± 0.53 |
DPC without SSC. DPC-GCN achieves higher means than Backbone on 11 of 13 datasets, with essentially unchanged performance on Physics and Questions. Mean gains on Cora, CiteSeer, Computer, Photo, WikiCS and Amazon-Ratings are approximately 0.76, 0.88, 0.46, 0.66, 0.73 and 1.21 percentage points, respectively. This comparison directly demonstrates degree-calibration gains without SSC.
DPC with SSC retained. Comparing DPC-GCN-S with Backbone + SSC, DPC still adds approximately 0.92, 0.54, 0.35 and 1.20 points on Photo, WikiCS, Computer and Amazon-Ratings, respectively. As you noted, the additional gains on Roman-Empire and Questions are essentially zero, and those on Minesweeper and Tolokers are small. Thus, DPC's benefit with SSC already present is dataset-dependent, with the four datasets above providing more direct positive examples.
SSC and the combined model. Comparing DPC-GCN-S with DPC-GCN, SSC improves mean performance on all five heterophilous datasets, including gains of approximately 1.00 and 0.83 points on Minesweeper and Questions; changes on homophilous datasets are only −0.20 to +0.06 points. These results distinguish the components' contributions: the backbone comparison supports DPC on its own, the SSC-retained comparison shows DPC's additional value on some tasks, and the larger gains on several heterophilous datasets primarily reflect SSC. The combined model's overall performance is interpreted through these component comparisons rather than attributed entirely to DPC.
3. Model-selection effort
Thank you for highlighting the importance of model-selection effort. Our tuning proceeds in stages with coordinate-wise refinement rather than a full Cartesian grid. For baselines, we prioritize released configurations that match the current evaluation setup and conduct supplementary searches when applicable settings are unavailable.
Taking Cora as an example, the corresponding DPC-GCN search stages consider hidden widths , depths , dropout rates , and coefficient initializations and . Backbone parameters are varied one axis at a time around , followed by separate initialization searches. DPC-GCN-S additionally considers SSC learning-rate multipliers at a fixed DPC configuration; layer-wise calibration coefficients are learned during training.
For comparison, Tuned-GCN follows the Classic GNN paper's Cora configuration: hidden width 512, three layers, dropout 0.7, learning rate 0.001, and 500 epochs, without normalization or residual connections. This setting incorporates the original study's tuning over hidden widths , depths , dropout rates , learning rates , and choices of normalization and residual connections. Reusing a released configuration therefore differs from using untuned defaults.
For baselines requiring adaptation to the current split or evaluation setup, we start from released configurations and perform supplementary searches. CPF on Cora retains its two-layer feature network and refines initial configurations coordinate-wise over hidden widths , filter orders , feature dropout , and filter dropout . Feature-network learning rates are , with additionally considered for the filtering branches. Weight decay covers for the feature network, for the structural-filter branch, and for the feature-filter branch.
H2GCN retains the public configuration's hidden width of 64 and two propagation rounds, and considers dropout , learning rates , weight decay , and ReLU/identity activations. These ranges describe the values considered in the respective search stages, not an exhaustive evaluation of all combinations.
The inherited tuning effort and local search scopes do differ across models. Appendices C.3–C.4 document the configuration sources, settings, and search ranges. We will further refine hyperparameter tuning during the rebuttal period by expanding candidates over relevant backbone and method-specific hyperparameters for baselines with limited coverage in existing configurations and supplementary searches. Selection will use validation performance to better assess their performance potential and the sensitivity of the current comparisons to the search scope.
4. Operator overhead at matched architecture
Using the five heterophilous datasets as examples, we measure the additional training time and GPU memory introduced by DPC and SSC under matched backbone configurations. Within each dataset, Backbone, DPC-GCN, and DPC-GCN-S share the same hidden width, depth, dropout, and training protocol, differing only in the calibration modules. Measurements use an exclusive RTX 5090. Each cell reports training time (ms/epoch) / peak allocated GPU memory (MiB), taking the median across three measurement blocks with rotated model order, each comprising 10 warmup and 50 measured training steps. Timing excludes graph preprocessing; memory includes the data, model, optimizer, cached graph, and training activations.
| Model | Ratings | Minesweeper | Roman-Empire | Questions | Tolokers |
|---|---|---|---|---|---|
| Backbone | 7.37 / 731.8 | 10.36 / 173.6 | 11.25 / 962.3 | 69.70 / 4450.7 | 4.48 / 200.8 |
| DPC-GCN | 8.49 / 739.1 | 19.75 / 189.5 | 13.44 / 974.0 | 76.71 / 4511.5 | 8.09 / 274.5 |
| DPC-GCN-S | 11.99 / 1029.7 | 24.28 / 260.6 | 16.95 / 1373.4 | 114.01 / 7074.6 | 10.28 / 329.2 |
Across these five datasets, DPC-GCN takes 1.10–1.91× the backbone's training time; adding SSC takes a further 1.23–1.49× relative to DPC-GCN. The first two rows isolate the overhead of DPC, and the last two isolate the additional overhead of SSC. These matched-architecture comparisons directly quantify the additional cost of calibration. Section 4.5 instead compares accuracy and resource use at each method's selected configuration, rather than claiming an operator-level speedup over the same GCN backbone; we will make this distinction explicit in the revision.
1. Local degree-coordinate effects
Thank you for the careful reading and independent check of Proposition 3. We agree that geometric non-redundancy, predictive utility, and architecture-level resource trade-offs are different claims. The additional evidence separates them.
We provide run-level differences and confidence intervals for all seven datasets in Table 6, focusing our analysis on Computer, WikiCS and Minesweeper as requested. We further examine how the local correction affects nodes with different structural discrepancies.
Run-level performance and uncertainty. The table compares full DPC with Graph-wide and Shuffled-centered. Each cell reports the mean difference; [five run-level differences]; [marginal 95% CI], in percentage points. Differences are DPC-GCN minus the control, so positive values favor DPC-GCN.
Runs are paired by their predefined run indices and corresponding data splits; 95% confidence intervals use paired t statistics on the run-level differences.
On the three datasets highlighted in your question, full DPC has higher mean performance than both controls. On Minesweeper, all five differences from Graph-wide are positive, with a mean gain of 0.415 points. On Computer and WikiCS, all five differences from Shuffled-centered are also positive, with mean gains of 0.225 and 0.339 points, respectively. The 95% paired interval excludes zero for Minesweeper versus Graph-wide; the remaining intervals include zero. These results show positive mean gains from the local coordinate on the highlighted tasks, with the clearest evidence on Minesweeper versus Graph-wide. Node-level effects. To further examine the contribution of the local correction, we hold all other model parameters fixed and remove it by setting λl=0 in every layer at inference. Test nodes are grouped by the structural discrepancy coordinate qi, without using labels or predictions; larger qi indicates a larger relative discrepancy between the node's normalized propagation rows. The table reports the mean increase in negative log-likelihood (NLL) after removal over five runs per dataset; positive values favor retaining the local term. Equal-qi nodes remain together, yielding two groups on Minesweeper.
Across all five runs on each of the three datasets, removing the local term while holding the other model parameters fixed increases mean NLL in the high-q group. This finding provides node-level evidence of the role of local calibration, revealing prediction responses associated with structural discrepancy beyond aggregate performance metrics. It also motivates further investigation of the relationship between local topology and calibration benefits.