- Rating
- 1
- Confidence
- 4
- Informativeness
- 1
- Presentation
- 1
- Soundness
- 1
Read review
Paper summary
The manuscript presents ER-GCN, a novel message-passing layer for Graph Neural Networks that mixes a standard GCN-like propagation operator with a dense operator derived from effective resistance theory. The paper claims that this hybrid layer removes structural zeros in the Jacobian that affect long-range propagation. Moreover, the authors propose a variational characterization of their methodology. Experiments explore Barbell topologies, TreeNeighborsMatch, and several real-world datasets (LRGB Peptides, TU, and ZINC).
Critical strengths
- I checked the theory and it is correct, proving that dense mixing layers are a sufficient condition to annihilate structural zeros in the Jacobian. Limitations are stated and acknowledged. I think the application of effective resistance in GNNs is an interesting topic, and the authors' implementation really is simple, neat, efficient and lightweight: it is as easy as modifying a GCN layer (excluding the pre-processing overhead), which I appreciate a lot.
- Experimental setup is broad and reveals, in full clarity, also the limitations of the method. I also found the ablation useful and detailed, and I think it is important to report relevant results in a complete and transparent way.
Critical weaknesses
- Limited novelty and evidence of long-range propagation. The main contribution of the manuscript is the construction of a dense node-mixing operator on top of standard GCN message passing. While the proposed theoretical results are sufficient to establish that such an operator will remove topologically introduced zeros (and produce positive Jacobians), they do not guarantee that information from distant nodes can still have a significant influence on a given node. In my opinion, the manuscript lacks a methodological and theoretical analysis that relates the elimination of structural zeros to meaningful message passing, that is, a proof of effective long-range propagation. Moreover, the main theoretical guarantees follow directly from the strictly positive off-diagonal entries of rather than from any specific property of ER; the assumptions in Section 3.2 only require . Uniform mixing would, indeed, satisfy the same properties. In fact, the authors include this comparison in an ablation study (see Table 8), where uniform stays competitive with the resistance-based kernel (uniform: vs. resistance: ), undermining the claimed benefits of resistance weighting. Finally, the chosen kernel assigns exponentially smaller weights to pairs of nodes with higher resistance than to lower-resistance pairs in the same row, suppressing signals from distant nodes. In conclusion, given the existing literature on long-range propagation and on effective resistance in GNNs, this contribution appears limited unless the manuscript is adequately revised and its long-range claims are addressed.
- Experimental evidence is limited. In addition to the aforementioned methodological limitations, I think that the experimental section presents problems that further weaken the manuscript's conclusions. After carefully reading Appendix C, I must point out that the reported score for each seed on the Barbell task is the lowest test MAE observed during training (manuscript lines 1401-1402, Table 5). In other words, the authors are selecting the best checkpoint on the test set, undermining the validity of the results. Secondly, the hyperparameter selection protocol is not documented. Appendix D thoroughly reports the selected hyperparameters, but neither the search space nor the selection criterion is described. This is particularly important given that the importance of hyperparameter selection and tuning in GNNs has already been analyzed in previous literature [5]. Besides the serious concerns about the tuning, another major experimental issue arises. The only long-range dataset experiment provided is Peptides from the LRGB suite. However, recent literature has extensively questioned the actual long-range nature of the task. In particular, in [2] the authors show substantial improvements from tuning MPNN baselines on LRGB; in [3] the authors formally question the actual need for long-range interactions to solve the Peptides task. Finally, results on ZINC and conventional graph datasets do not establish, nor prove, improved long-range propagation arising from the method.
I must also point out major clarity issues. The manuscript formulation and writing style is convoluted and not ready for publication: core math concepts are introduced after their definition (just for example, the matrix is introduced in sec. 3, but its structure only defined in sec. 4; notation is hard to follow; terms like "structural zeros" or are never defined in the text).
References:
-
[1] Velingker et al. Affinity-Aware Graph Networks. NeurIPS, 2023.
-
[2] Tönshoff, J., et al. Where Did the Gap Go? Reassessing the Long-Range Graph Benchmark. arXiv preprint arXiv:2309.00367, 2023.
-
[3] Bamberger, J. et al. On Measuring Long-Range Interactions in Graph Neural Networks. ICML, 2025.
-
[4] Mathys, J. et al. LRIM: a Physics-Based Benchmark for Provably Evaluating Long-Range Capabilities in Graph Learning. ICLR, 2026.
-
[5] Luo, Y. et al. Classic GNNs are Strong Baselines: Reassessing GNNs for Node Classification. arXiv preprint arXiv:2406.08993, 2024.
-
[6] Gravina et al. Anti-Symmetric DGN: a stable architecture for Deep Graph Networks. arXiv preprint arXiv:2210.09789, 2023.
Questions for authors
- Can the authors clarify the positioning of their methodology with respect to existing ER literature in GNNs (e.g., [1])?
- Can the authors provide a quantitative and methodological analysis of the actual long-range capabilities of the ER-GCN layer, with particular regard to the weaknesses pointed out in W1?
- Can the authors address the experimental limitations listed in W2? In particular, can the authors provide empirical long-range evidence beyond the discussed LRGB results, and test their method on long-range datasets (see LRIM [4] as a solid dataset) and against long-range MPNNs (see A-DGN [6] as an example of a solid long-range baseline)? Can they provide a real long-range sensitivity analysis according to [3]?
- Again on the experimental setup, can the authors provide the hyperparameter search space used to train ER-GCN?
8 comments
2. Quantitative & Methodological Analysis of Long-Range Propagation(Q2, Q3, W1)
To provide evidence of ER-GCN's task-meaningful message passing, we conducted a four-level quantitative, range [3], and causal analysis on trained Barbell 10-layer ER-GCN models:
(A) Layer-Level Operator Attenuation & Jacobian Factorization
In the hybrid layer , the row-wise weight ratio is .
- For target nodes to distant source nodes , entries are (hop 11, ), (hop 12, ), and (hop 13, ).
- Summing over the 12 source-clique nodes, receives of a resistance- row. With learned ( to ), direct per-layer mixed mass beyond 10 hops is per layer.
- Factoring the exact 1-layer Jacobian as \left\\|\frac{\partial H_u^{l+1}}{\partial H_v^l}\right\\|_{1,\mathrm{entry}} = (M_l)_{uv} C_{u,l}(X) where , the trained network actively amplifies early-layer signals (, yielding direct 1-layer Jacobian magnitudes of at hops 11, 12, 13) before projecting down at the readout layer ().
(B) Long-Range Sensitivity Analysis Following Bamberger et al. [3]
Using the normalized hop sensitivity distribution where T_X(u,h) = \sum_{v: d_G(u,v)=h} \left\\|\frac{\partial \hat{y}_u}{\partial X_v}\right\\|_1, we compute the influence-weighted range [3], the -mass radius , and the sensitivity fraction beyond the 10-hop barrier ():
For ER-GCN with and , it has full-input range hops ( CI ), with () of total sensitivity strictly beyond 10 hops and -mass radius hops (matching the oracle target distance of hops; source-restricted range hops). Individual source–target pair influences remain strong across all distant shells: (hop 11), (hop 12), and (hop 13).
(C) End-to-End Task Alignment & Causal Interventions
To verify that this distant Jacobian mass reflects faithful message passing rather than noise, we test against the ground-truth generator (oracle gradient at hops 11–13):
- Directional Gradient Alignment: Cosine similarity between ER-GCN's input gradient and the oracle gradient is (), specifically , , and at hops 11, 12, and 13. ER-GCN recovers the exact projection vector across the 11–13 hop bottleneck.
- Causal Source-Resampling Interventions: Independently resampling features (11–13 hops away) on held-out graphs while freezing all other inputs/weights yields prediction shifts tracking with correlation (), regression slope is (ideal ). Single-node interventions at hops 11, 12, and 13 individually achieve correlations of .
(D) Bandwidth () Sweep
We directly quantifies the reviewer's concern about exponential decay. When is much smaller than , exponential decay indeed suppresses ( row mass, , MAE ). However, once matches the resistance scale (), attenuation is mild: reaches (), (, MAE ), and (), preserving causal transfer ().
3. Why Exponential Resistance Kernel? (W1)
We choose because it uniquely solves the entropy-regularized distortion minimization (Section 4, Eq. 13 & Lemma A.2): \boldsymbol{q}_u^\star = \arg\min_{\boldsymbol{q}_u \in \Delta_u} \left\\{ \sum_{v \neq u} q_{uv} R_{\mathrm{eff}}(u,v) + \sigma\\, \mathrm{KL}(\boldsymbol{q}_u \\| \boldsymbol{\pi}_u) \right\\}, \quad \text{where } \left\\|\overline{\boldsymbol{h}}_u(\boldsymbol{q}_u) - \boldsymbol{h}_u\right\\|_2^2 \le \operatorname{tr}(\boldsymbol{H}^\top \boldsymbol{L} \boldsymbol{H}) \sum_{v \neq u} q_{uv} R_{\mathrm{eff}}(u,v). Uniform mixing (, corresponding to ) minimizes only entropy while ignoring the topological distortion bound , indiscriminately mixing unrelated graph regions when all nodes carry active features. Conversely, collapses support. For finite , the exponential kernel uniquely balances full global support () with bounded geometric distortion.
4. Protocol Fix, New Baselines, and LRIM Evaluation (Q3, W2)
(a) Barbell Checkpoint Selection Fix: We thank the reviewer for pointing out this issue. We added an independent validation split for checkpoint selection and updated all Barbell results in Table 1.
(b) Following the reviewer's suggestion, we have tested A-DGN [4] on Barbell (MAE ) and TreeMatch (Acc % ). Experiments on other datasets are in progress and will be included in the revision
| Model | ||||||
|---|---|---|---|---|---|---|
| A-DGN [4] | N/A | N/A | N/A | 99.70±0.51 | 15.30±2.10 | 16.00±0.00 |
| ER-GCN | 0.1334±0.0275 | 0.1396±0.0269 | 0.1970±0.0214 | 100.00±0.00 | 99.70±0.67 | 99.65±1.11 |
A-DGN is 1-hop local and therefore N/A on Barbell . On TreeMatch, A-DGN suffers over-squashing (16.0% at ) while ER-GCN keeps Acc and lowest MAE.
(c) We agree with [3, 5] regarding Peptides. We thank the reviewer for recommending LRIM [6], an exemplary, physics-grounded benchmark where Ising-system targets provably depend on long-range interactions across graphs with dense node features, paired with explicit complexity reporting.
To directly test overcoming the hard receptive-field barrier, we evaluated ER-GCN and the LRIM baselines following LRIM's official hyperparameter selection protocol but at fixed depth (well below lattice diameter).
| Model | LRIM-16-hard | LRIM-32-hard | LRIM-16-easy | LRIM-32-easy |
|---|---|---|---|---|
| GIN [7] | -1.6795±0.0112 | -1.2213±0.0089 | -2.5964±0.0177 | -2.2461±0.0161 |
| GatedGCN [8] | -1.6006±0.0078 | -1.2049±0.0314 | -2.6478±0.0139 | -2.1303±0.0249 |
| GatedGCN-VN$_G$ [9] | -1.7315±0.0149 | -1.2185±0.0037 | -2.6887±0.0071 | -2.2200±0.0055 |
| ER-GCN | -1.7012±0.0657 | -1.2245±0.0702 | -2.8528±0.0430 | -2.3412±0.0578 |
ER-GCN beats GIN/GatedGCN on all splits and ranks 1st on 3/4 splits, validating ER decay on dense active features. GPS [10] and its variants on other LRIM datasets under the same protocol are currently in progress and will be included and cited upon acceptance.
5. Hyperparameter Search Space & Protocol (Q4, W2)
We have add the hyperparameter search space to Appendix D.
6. Manuscript Organization and Presentation of (W3)
We agree with this critique. We have reorganized it and moved the definition of and its explanation to their proper place for better readability and logical progression.
We sincerely thank the reviewer for their feedback and their thoughtful evaluation. Below we address all points.
1. Positioning w.r.t. Existing Effective-Resistance Literature in GNNs (Q1, W1)
Prior work incorporates effective resistance (ER) into GNNs in roles fundamentally distinct from ER-GCN: