- Rating
- 3
- Confidence
- 2
- Informativeness
- 2
- Presentation
- 3
- Soundness
- 2
Read review
Paper summary
This work studies unsupervised multi-view graph domain adaptation for cross-domain node classification. The authors propose Cross-Domain Aggregated Class Distance (CD-ACD), a metric that combines cross-domain same-class alignment distances with intra-domain and cross-domain inter-class separation distances to capture class-level structural discrepancies. Building on this metric, the authors develop the Class Centroid-Guided Alignment (CCGA) framework, which dynamically assigns weights to different graph views and integrates graph feature reconstruction, contrastive alignment, source-domain classification, and target-domain entropy minimization. Experiments on the ACM and MAG datasets show that the approach improves classification accuracy and Macro-F1 compared with existing baselines, including SAGA. Ablation studies and hyperparameter sensitivity analyses are also conducted to evaluate the framework.
Critical strengths
- This paper tackles a significant challenge in multi-view graph domain adaptation by accounting for class-level structural discrepancies that view-level distribution alignment may fail to capture sufficiently. Incorporating class centroid information into the alignment process provides an intuitive way to enhance cross-domain representation learning.
- The proposed method delivers competitive results across multiple transfer tasks on the ACM and MAG datasets, improving both classification accuracy and Macro-F1 relative to existing methods. The experiments also include ablation studies and hyperparameter analyses, offering preliminary evidence for the effectiveness of the proposed framework.
Critical weaknesses
- Equation (3) combines cross-domain same-class distances with negatively weighted inter-class separation distances. A lower CD-ACD value may therefore reflect greater class separation as well as closer cross-domain alignment. Both can be useful for adaptation, but this does not by itself establish that CD-ACD reliably measures class-level structural shift. The validation should distinguish these effects more clearly. In addition, although the authors claim to dynamically align individual class pairs, Equations (7)-(9) mainly introduce view-level weighting. How the aggregated class distances lead to explicit alignment of individual class pairs remains unclear. These points leave the interpretation of CD-ACD and its role in class-level alignment insufficiently supported.
- Equation (8) (p. 6) is described as a bidirectional InfoNCE contrastive loss, but it does not specify positive pairs or normalize their similarity against competing pairs. As written, the loss can be minimized by mapping all nodes to the same nonzero projected representation, making every cross-domain cosine similarity equal to one. This is a concern about the stated objective, rather than evidence that collapse occurred in the experiments. Since reconstruction and classification operate on the encoder representations, it is unclear what prevents collapse in the projection head. Clarifying whether Equation (8) matches the implementation is necessary to assess the alignment objective.
Questions for authors
- Given that Equation (3) combines cross-domain same-class distances with negatively weighted intra-domain and cross-domain inter-class separation distances, so that a lower CD-ACD value may result from increased inter-class separation rather than reduced cross-domain structural discrepancy, can CD-ACD reliably measure the class-level structural shift it is intended to characterize?
- Although the authors claim to dynamically align individual class pairs, Equations (7)-(9) primarily introduce view-level weighting. How is class-pair-specific alignment explicitly achieved, and does this weaken the justification for the paper’s central methodological contribution?
- Equation (8) is described as a bidirectional InfoNCE contrastive loss, but its formulation does not explicitly distinguish positive from negative sample pairs or include the standard normalization between positive-pair similarity and competing similarities. Does this objective therefore merely encourage cross-domain representational similarity without explicitly preserving class discrimination, raising concerns about the correctness of the proposed alignment objective?
- Appendix C states that the experiments were repeated five times, but the main results report only single performance values without standard deviations or other measures of variability. Can the statistical reliability of the reported improvements be assessed?
6 comments
Official Comment by Authors (2/4)
Response to W2 and Q3: The Bidirectional InfoNCE Contrastive Loss and Class Discriminability
Thank you for your detailed comments on the bidirectional InfoNCE contrastive loss and its class discriminability.
The bidirectional InfoNCE contrastive loss presented in the paper does not explicitly construct same-class positive pairs, cross-class negative pairs, or class-dependent masks. The original loss is given in Section 4.2, Eq. (8) (p. 6, lines 275–281). To investigate class discriminability during actual training, we analyze distances between different classes in the prediction space and provide the full explanation in Appendix K (CLASS DISCRIMINABILITY ANALYSIS) (p. 27–28, lines 1406–1510; Tables 4–7). Selected results are shown in Table 1. Different classes retain distinguishable distances in the prediction space, indicating that the final predictions have not degenerated into identical outputs across all classes and retain class discriminability.
Table 1. Relative class representation distances in the prediction space on CN→DE under view 1 with the original loss.
| S/T | C0 | C1 | C2 | C3 | C4 | C5 | C6 | C7 | C8 | C9 |
|---|---|---|---|---|---|---|---|---|---|---|
| C0 | 1.00 | 5.12 | 5.80 | 2.48 | 4.55 | 4.57 | 3.38 | 6.01 | 1.57 | 1.59 |
| C1 | 4.55 | 1.00 | 2.13 | 5.17 | 2.57 | 2.38 | 5.63 | 2.28 | 4.70 | 4.59 |
| C2 | 4.96 | 2.00 | 1.00 | 5.51 | 2.41 | 2.21 | 5.94 | 1.25 | 5.06 | 4.98 |
| C3 | 2.76 | 5.32 | 5.89 | 1.00 | 4.95 | 4.98 | 1.74 | 6.00 | 3.05 | 3.33 |
| C4 | 2.89 | 1.55 | 1.63 | 3.37 | 1.00 | 0.79 | 3.75 | 1.90 | 2.96 | 2.91 |
| C5 | 3.80 | 2.03 | 2.20 | 4.41 | 1.40 | 1.00 | 4.91 | 2.55 | 3.89 | 3.82 |
| C6 | 2.59 | 5.11 | 5.58 | 0.90 | 4.69 | 4.70 | 1.00 | 5.70 | 2.81 | 3.13 |
| C7 | 4.14 | 1.73 | 1.04 | 4.59 | 2.23 | 2.07 | 4.93 | 1.00 | 4.26 | 4.18 |
| C8 | 1.24 | 7.02 | 8.02 | 3.67 | 6.30 | 6.29 | 4.95 | 8.41 | 1.00 | 1.24 |
| C9 | 0.71 | 4.16 | 4.74 | 2.16 | 3.71 | 3.72 | 2.88 | 4.97 | 1.12 | 1.00 |
Official Comment by Authors (3/4)
Response to W2 and Q3: The Bidirectional InfoNCE Contrastive Loss and Class Discriminability
To further investigate the effect of explicitly constructing same-class positive pairs and cross-class negative pairs with class-dependent masks, we design a bidirectional InfoNCE contrastive loss incorporating class encoding.
We denote this variant with explicit class-aware positive and negative pairs as CCGA-CP and compare its class discriminability and classification performance with CCGA, which uses the original loss presented in the paper. The full experimental description has been added to Appendix L (COMPARISON OF BIDIRECTIONAL ALIGNMENT LOSSES) (p. 29–31, lines 1514–1634; Table 8), with selected results shown in Tables 2 and 3. These correspond to Table 10 (p. 30, lines 1589–1602) and Table 8 (p. 29, lines 1550–1565), respectively, in Appendix L.
The results show that explicit positive and negative sample constraints improve the measured class discriminability but do not yield corresponding overall improvements in Accuracy or F1. This indicates that stronger explicit class contrastive constraints do not necessarily translate into better target-domain classification performance under the current experimental settings, and may instead introduce additional computational overhead.
Table 2. Relative class representation distances in the prediction space on CN→DE under view 1 with the new loss.
| S/T | C0 | C1 | C2 | C3 | C4 | C5 | C6 | C7 | C8 | C9 |
|---|---|---|---|---|---|---|---|---|---|---|
| C0 | 1.00 | 14.31 | 12.37 | 10.19 | 10.10 | 10.36 | 13.97 | 12.72 | 1.15 | 1.24 |
| C1 | 8.34 | 1.00 | 5.94 | 6.85 | 6.97 | 6.94 | 7.68 | 5.92 | 8.46 | 8.38 |
| C2 | 144.67 | 113.77 | 1.00 | 189.96 | 40.30 | 37.10 | 234.06 | 4.75 | 145.87 | 143.17 |
| C3 | 8.76 | 10.50 | 12.92 | 1.00 | 12.18 | 12.33 | 2.72 | 13.11 | 8.78 | 9.01 |
| C4 | 25.88 | 29.71 | 7.90 | 39.66 | 1.00 | 0.29 | 50.06 | 8.83 | 26.01 | 25.48 |
| C5 | 76.99 | 87.87 | 23.04 | 117.70 | 3.31 | 1.00 | 148.52 | 25.82 | 77.38 | 75.80 |
| C6 | 51.36 | 49.92 | 66.73 | 14.45 | 64.95 | 65.55 | 1.00 | 67.43 | 51.59 | 52.48 |
| C7 | 58.46 | 43.71 | 2.14 | 75.37 | 17.86 | 16.61 | 92.49 | 1.00 | 58.96 | 57.88 |
| C8 | 1.24 | 30.48 | 24.48 | 22.71 | 19.16 | 19.76 | 31.25 | 25.26 | 1.00 | 1.40 |
| C9 | 0.72 | 22.62 | 18.92 | 16.85 | 15.19 | 15.62 | 23.10 | 19.50 | 1.19 | 1.00 |
Table 3. Experimental results of CCGA with different loss functions (%).
| Method | Metric | A1→A2 | A2→A1 | CN→US | US→CN | JP→CN | CN→JP |
|---|---|---|---|---|---|---|---|
| CCGA | Acc | 60.4±0.35 | 53.8±2.57 | 61.2±1.08 | 54.2±0.54 | 55.5±0.33 | 59.4±0.22 |
| CCGA | F1 | 45.7±0.39 | 53.3±2.56 | 43.6±0.50 | 43.8±0.30 | 43.8±0.65 | 47.5±0.30 |
| CCGA-CP | Acc | 60.5±0.40 | 53.6±6.87 | 61.1±0.65 | 53.9±0.22 | 55.4±0.44 | 59.4±0.13 |
| CCGA-CP | F1 | 45.5±0.33 | 54.0±6.87 | 43.9±1.24 | 44.1±0.37 | 43.9±0.51 | 47.7±0.29 |
| Method | Metric | DE→FR | FR→DE | RU→US | US→RU | DE→CN | CN→DE |
|---|---|---|---|---|---|---|---|
| CCGA | Acc | 55.8±0.24 | 57.0±0.26 | 42.4±0.46 | 45.9±0.24 | 47.3±0.73 | 42.6±0.29 |
| CCGA | F1 | 50.0±0.12 | 50.4±0.68 | 41.9±0.63 | 41.9±0.25 | 43.1±1.04 | 41.5±0.21 |
| CCGA-CP | Acc | 55.6±0.31 | 57.1±0.11 | 42.3±1.24 | 45.9±0.19 | 47.1±0.29 | 42.7±0.24 |
| CCGA-CP | F1 | 50.1±0.17 | 50.5±0.16 | 42.0±1.11 | 41.8±0.21 | 43.2±0.36 | 41.6±0.28 |
Official Comment by Authors (1/4)
Thank you for your insightful comments and constructive suggestions. We address your concerns and questions point by point below.
Response to the Shared Concern in W1 and Q1: Effectiveness of CD-ACD in Measuring Class-Level Structure
Thank you for your insightful comments on whether CD-ACD effectively measures differences in class-level structure.
To further investigate the relationship between CD-ACD and class-level training dynamics, we have added an analysis in Appendix F (EXPERIMENT DETAIL OF EMPIRICAL STUDY) (p. 18, lines 953–960; pp. 18–19, lines 970–974; Figure 3, p. 19, lines 976–1024). For each class, we compute the difference between the source-domain supervised cross-entropy loss and the target-domain prediction entropy, LS−LT, and plot its evolution alongside the minimum CD-ACD score over training epochs. The definitions of these losses are provided in Section 4.3, Eqs. (10)–(11) (p. 6, lines 297–315).
Although the class-specific curves exhibit local fluctuations, they generally decrease during training, accompanied by a corresponding downward trend in the minimum CD-ACD score. This consistency in the observed trends across multiple classes provides supporting evidence that CD-ACD reflects class-level structural information.
In addition, the ablation experiments on CD-ACD components in Appendix G (MODEL ABLATION STUDY) (p. 20, lines 1054–1079; Figure 5, p. 21, lines 1080–1133; Figure 6, p. 22, lines 1134–1174) assess, to some extent, the contributions of the same-class alignment distance and inter-class separation distances to the final model performance.
Table. Ablation results of individual CD-ACD components (%). Each cell reports Accuracy / F1.
CCGA₁ removes the intra-domain class separation distance; CCGA₂ removes the cross-domain cross-class separation distance; CCGA₄ removes the cross-domain same-class alignment distance.
Response to the Shared Concern in W1 and Q2: How CD-ACD Guides Alignment
Thank you for this question and for pointing out the need to further clarify how CD-ACD uses class-level structural information to guide subsequent alignment.
CD-ACD is computed from source-domain and target-domain class centroids. It integrates information about same-class and cross-class pairs within and across domains to select the cross-domain alignment view and generate multi-view consistency weights. In this way, structural information from different classes is aggregated to jointly guide the cross-domain alignment and intra-domain multi-view consistency terms in Eq. (9). The class-centroid construction and score are defined in Section 3 (p. 4, lines 165–198), and the alignment mechanism is described in Section 4.2 (p. 5, lines 261–269; p. 6, lines 270–291).
Cross-domain alignment. The model selects the view k with the lowest CD-ACD score and uses its corresponding weight ωb to scale the bidirectional alignment loss. Consequently, cross-domain representation alignment is jointly guided by same-class correspondence, cross-class separation, and intra-domain class separability. See Section 4.2 (p. 6, lines 270–281 and 287–291).
Intra-domain multi-view consistency. Within the source and target domains, the model constrains differences in second-order representation statistics between each view and the baseline view. The weights ωv determine the contributions of individual views, while 1−ωb controls the overall weight of this consistency term. Thus, intra-domain consistency is guided not only by statistical differences between views but also by class-level structural information. See Section 4.2 (p. 6, lines 284–293).