- Rating
- 3
- Confidence
- 4
- Informativeness
- 2
- Presentation
- 2
- Soundness
- 2
Read review
Paper summary
The paper focuses on the fact that existing text-attributed graph (TAG) methods suffer from two issues: (1) structural-semantic mismatch arising from divergent modality distributions, and (2) heterogeneous local topologies that a single geometric space cannot adequately capture. To address these issues, this paper proposes Riemannian Multi-Expert Structural-Semantic Decoupling (R-MESD) to separate structural and semantic representations. The structural branch employs adaptive Hyperbolic and Spherical experts, with an LLM-based structural description for calculating node-level routing weights. The semantic branch encodes textual information as tangent-space vectors anchored at learned structural positions, and further applies cross-geometry semantic alignment and tangent-space semantic attention. Experiments on five datasets show improvements over existing GNNs and TAG baselines.
Critical strengths
- The paper tackles a practical problem where structural-semantic mismatch arises from divergent modality distributions in text-attributed graph learning.
- The proposed framework is technically sound and includes meaningful experimental analysis, such as the performance comparison and the module ablation study.
Critical weaknesses
None
Questions for authors
- Figure 1 and Table 4 discuss the model performance w.r.t. heterophily intervals, this makes me doubt whether advanced heterogeneous graph models can also show good performance. So the authors are supposed to provide comparisons with advanced heterogeneous graph models.
- The training objective in Eq. (32) doesn’t include a classification loss or a loss using the labels. Is R-MESE an unsupervised model? If so, it should be compared with more unsupervised methods, not only supervised methods.
- The ablation should provide the standard deviations.
- Why do the authors only use two geometric branches in Table 3, but not three branches with H, S, and E?
- The function of structural positions needs more explanation. Why is it useful to be concatenated with the semantic representations in Eq. (28)?
- In Section 4.4, the authors state that R-MESD yields clearer class separation, indicating more discriminative node representations. However, I just find little difference between (b) and (c).
- I wonder whether the promising model performance is attributed to the powerful LLM. Do the authors use the same LLM as the baseline models or a more powerful one?
3 comments
Q7: Our gains do not rely solely on the LLM. Qwen3-8B has fewer parameters than the baselines’ LLMs. To examine its role, we compare two routing variants without LLM descriptions. Structure-only routing uses structural statistics, while Direct-feature routing also uses text embeddings. As shown in Table 4, both perform well, showing that our model remains effective without description generation.
Table 4. Comparison of Routing Inputs on Accuracy (%).
| Methods | Cora | CiteSeer | PubMed | ArXiv-2023 | OGBN-ArXiv |
|---|---|---|---|---|---|
| Structure-only routing | 87.76 ± 0.74 | 80.81 ± 1.36 | 88.06 ± 0.57 | 73.09 ± 1.92 | 72.22 ± 1.08 |
| Direct-feature routing | 89.33 ± 0.63 | 79.81 ± 2.19 | 91.06 ± 0.40 | 74.59 ± 1.83 | 74.22 ± 1.35 |
| R-MESD | 92.37 ± 0.31 | 81.65 ± 1.17 | 92.59 ± 0.62 | 77.98 ± 1.62 | 76.11 ± 1.45 |
Thanks for the authors, all my issues have been addressed. I will maintain my score.
Q1: Following your suggestion, we compare our method with two advanced heterophilic graph models, H2GCN and GPR-GNN. As shown in Table 1, our method outperforms both models on all five datasets, confirming its advantage over methods designed for heterophilic graphs.
Table 1. Comparison with Heterophily-Oriented Methods on Accuracy (%).
Q2: R-MESD uses unsupervised encoder pretraining followed by supervised downstream classification. The pretraining objective in Eq. (32) uses no node labels. After pretraining, we freeze the encoder and train a downstream GNN classifier using the training labels.
Following your suggestion, we include DGI, GraphMAE2, and BGRL as additional unsupervised methods. As shown in Table 2, our method achieves higher accuracy on all five datasets than these unsupervised methods.
Table 2. Comparison with Unsupervised Methods on Accuracy (%).
Q3: Following your suggestion, we report the mean accuracy and standard deviation for the ablation experiments.
Table 3. Ablation Results on Accuracy (%).
Q4: Our R-MESD omits a separate Euclidean branch because hyperbolic and spherical experts with curvature near zero can approximate Euclidean geometry. Experts with different curvatures thus capture both curved and nearly Euclidean structures. Table 3 in our paper further supports this choice: H/S outperforms configurations with a Euclidean branch.
Q5: Concatenating structural positions in Eq. (28) gives the classifier direct access to topology, helping distinguish nodes with similar texts but different structures. It also preserves structural features as separate inputs, allowing the classifier to learn their importance for the task.
Q6: There seems to be a misunderstanding about the comparison. The clearer class separation refers to the improvement of both (b) and (c) over the initial representations in (a), rather than a large difference between (b) and (c).