- Rating
- 2
- Confidence
- 3
- Informativeness
- 2
- Presentation
- 2(−1)
- Soundness
- 2
Read review
Paper summary
This paper addresses domain generalization in vision-language models. It argues that standard single-text prototype class representations in CLIP are limited under domain shift. To solve this, the authors propose representing each class as a low-dimensional semantic subspace and classifying images by measuring their alignment with both the class centroid and its subspace. Across five DomainBed benchmarks, the approach achieves consistent improvements over CLIP prompt-learning baselines. Aablation study shows that the performance gain comes from the subspace structure itself, rather than simply using multiple prompts.
Critical strengths
- The idea is simple and easy to follow. Instead of collapsing all the text descriptions of a class into a single prototype, the paper keeps their spread as a low-rank subspace, and scores an image by both its distance to the class center and its fit to the subspace. The write-up is clear, and the method looks straightforward to reproduce.
- Tables 3 and 4 provide controlled comparisons that isolate the effect of the subspace structure. With the same 30 descriptions per class, averaging or taking the max over prompts performs worse than organizing them into a subspace (Table 3), and replacing each class's basis with another class's basis reduces accuracy (Table 4). These results indicate that the improvement is tied to the subspace structure and to class-specific directions, rather than to the use of multiple prompts alone.
Critical weaknesses
- The related work is incomplete and the main concern is that the core idea is not new. The paper situates itself only within CLIP-based prompt learning, and does not discuss the line of DG research that likewise models classes or features through subspaces and variation directions ([1] ISR, ICML 2022; [2] FVD, ACM MM 2021), nor recent CLIP-side work using diverse text prompts for DG ([3] Wen et al., CVPR 2025). Since the idea of "a class is a set of variation directions rather than a point" has precedents, a discussion of these works seems necessary to clarify what is new here.
- Justification of key design choices is thin. is chosen without showing the singular-value spectrum (e.g., cumulative energy), which is a natural and expected analysis for an SVD-based method. In addition, the loss weights (, ) and the adversarial radius have no reported sensitivity analysis, while Table 2 shows each auxiliary objective contributes within roughly one point, making it difficult to judge how stable these contributions are.
- The method depends on descriptions from a single LLM. The prompt bank comes entirely from GPT-5.6, and that is the only source ever tested. Since the whole method is built on top of these descriptions, I wonder how much the generator matters: would descriptions from a smaller or open-source LLM give similar subspaces and results? And what happens if some descriptions are wrong or repetitive? At this point it is hard to tell whether the gains come from the general idea of using multiple descriptions, or specifically from what GPT-5.6 knows. There is also a practical reproducibility concern, since GPT-5.6 is proprietary and its outputs may change over time — releasing the full prompt bank helps, but the pipeline itself cannot be re-run independently
[1] Provable Domain Generalization via Invariant-Feature Subspace Recovery. ICML 2022
[2] Domain Generalization via Feature Variation Decorrelation. ACM MM 2021
[3] Domain Generalization in CLIP via Learning with Diverse Text Prompts. CVPR 2025
Questions for authors
- What do the per-class singular-value spectra look like, and do they support r=8? How does performance change with λ_res, λ_inv (e.g., ±10×), and ε?
- How sensitive is the method to the description source? Would descriptions from a different LLM or human-written attributes yield similar subspaces and performance?
- The paper's central claim is that the textual subspaces support domain-invariant prediction, but I could not find direct evidence for this, e.g., a visualization or measurement showing that unseen-domain features lie closer to the class subspaces than a point-prototype baseline. Is there a way to demonstrate this, even on a toy setup?
2 comments
W3/Q2
Thank you for raising concerns about description sources and quality. We conducted additional cross-generator experiments and controlled tests with repeated and incorrect descriptions on OfficeHome. All settings use descriptions per class, rank , and the same training and evaluation protocol.
We independently generated description banks using DeepSeek-V4.1-Flash-1017 and GPT-6.1. For description quality, the duplicate setting replaces six descriptions per class with copies of remaining same-class descriptions; the incorrect setting replaces them with descriptions from other classes. Replacement rules and the random seed are fixed.
| Description bank | Accuracy |
|---|---|
| Original GPT-5.6 | |
| DeepSeek-V4.1-Flash-1017 | |
| GPT-6.1 | |
| Original bank, duplicate descriptions | |
| Original bank, incorrect-class descriptions |
Changing generators reduces accuracy by and percentage points, respectively. Performance remains relatively close across the tested sources, indicating that the method is not effective only with the original generator. This does not imply identical subspace geometry across generators.
Duplicate and incorrect descriptions reduce accuracy by and percentage points, respectively. Duplicates reduce diversity and change the statistical weighting of descriptions, whereas incorrect descriptions introduce class-mismatched information. These results indicate some tolerance to the tested corruption, while confirming the importance of description quality. The incorrect setting is synthetic wrong-class contamination, not an estimate of the error rate in naturally generated descriptions.
In practice, errors and redundancy can be reduced through explicit generation instructions requiring correct class identity, observable attributes, and distinct descriptions; duplicate and semantic-similarity checks followed by regeneration; and independent-model cross-checks of class relevance and attribute plausibility, with manual inspection of questionable descriptions. Replacements should preserve the description count per class. These measures reduce risk rather than guarantee error-free banks.
We will include these findings and practical recommendations in the revised manuscript and clarify the method's scope with respect to description sources and quality.
Q3
Thank you for suggesting direct evidence of unseen-domain alignment. We conducted inference-time diagnostics on all four held-out OfficeHome domains. For each domain, we fixed the trained model, image features, and class centers, comparing correct rank- bases with shuffled bases of the same rank. Shuffled results are averaged over permutations.
We measure squared distance to the ground-truth class affine subspace and the classification margin, defined as the ground-truth score minus the highest incorrect-class score. Lower distance and higher margin are preferable.
| Target domain | Correct-basis distance ↓ | Shuffled-basis distance ↓ | Correct-basis margin ↑ | Shuffled-basis margin ↑ |
|---|---|---|---|---|
| Art | 1.0275 | 1.2171 | 0.1261 | 0.1042 |
| Clipart | 1.0686 | 1.0904 | 0.1112 | 0.0867 |
| Product | 1.0356 | 1.0772 | 0.1594 | 0.1381 |
| Real World | 0.9192 | 0.9783 | 0.1858 | 0.1621 |
| Equally weighted mean | 1.0127 | 1.0908 | 0.1456 | 0.1228 |
Correct bases consistently yield smaller distances and larger margins, supporting both geometric alignment and discriminative value on unseen domains.
For the point-prototype comparison, the class center lies within its affine subspace, so subspace distance cannot exceed center distance. Smaller distance alone therefore does not establish the usefulness of class-specific directions. Our equal-rank, shared-center shuffled control specifically tests these directions.
The complementary training ablation in Table 5 gives for prototype-only classification, for projection-only classification, and for their combination. Together, these results show that subspace information supplements the prototype and that correct textual directions improve unseen-domain alignment and classification margins.
We will include these diagnostics while limiting our conclusion to their demonstrated geometric and discriminative value, rather than claiming complete feature-distribution invariance across domains.
W1
Thank you for pointing out these omissions. We agree that subspace-based feature modeling has precedents. We will discuss ISR, FVD, and Wen et al., and position our contribution as class-description-derived affine geometry for classification and orthogonal prediction-consistency training.
ISR recovers invariant-feature subspaces from first- or second-order class-conditional statistics across training environments, with guarantees under its data model. We instead estimate class centers and bases from textual descriptions for image–class matching and local perturbation constraints. Our objective therefore differs from invariant-feature recovery.
FVD separates variation in visual feature space and applies class-decorrelation regularization to mitigate negative transfer. We retain principal textual variation directions as part of class representations, use affine projections for classification, and constrain predictions under bounded orthogonal perturbations. The direction sources, classification roles, and training constraints differ.
Wen et al. use domain-simulating prompts for feature suppression, consistency, and diversification. Our class-specific textual geometry instead jointly defines the affine classifier and orthogonal consistency perturbations, connecting representation and stability training through the same bases.
Controlled comparisons support these designs. With identical descriptions and a shared training objective in Table 3, our subspace outperforms maximum similarity by 1.1, 0.9, and 0.9 percentage points on TerraIncognita, VLCS, and OfficeHome. The additional same-rank linear control achieves 84.27±0.18% on OfficeHome, versus 85.2±0.1% for the affine construction, a gain of 0.93 percentage points. Table 4 retains class centers and rank while shuffling class-to-basis correspondence during training and inference, reducing VLCS and OfficeHome accuracy by 3.39 and 2.45 percentage points. We will clarify how these controls support our incremental contributions relative to prior subspace-based approaches.
W2/Q1
Thank you for identifying the insufficient justification of rank selection and hyperparameter sensitivity. We supplemented the paper with class-wise textual singular-value spectra and sensitivity analyses of the two auxiliary loss weights and adversarial perturbation radius on OfficeHome.
For rank selection, we computed the singular-value spectrum of centered description embeddings from the frozen CLIP text encoder for each of the 65 OfficeHome classes. Cumulative energy is defined as the sum of squared leading singular values divided by the sum of all squared singular values. With 30 descriptions per class, the first 8 directions retain a median energy of 69.54%, with an interquartile range of 68.71%–70.88% and a class-wise range of 66.08%–73.84%. Thus, rank r=8 retains a relatively consistent proportion of the principal textual variation across classes while truncating the remaining directions.
Energy retention alone does not determine classification performance. We therefore interpret these spectra alongside the existing rank ablation: increasing rank retains more textual variation but does not necessarily improve generalization. Together, the classification results and spectral analysis support r=8 as an effective empirical trade-off. We will include the per-class spectra and cumulative-energy plots in the revised manuscript.
For hyperparameter sensitivity, we conducted additional single-run analyses on OfficeHome, changing one parameter at a time while keeping all other settings fixed. We tested one-tenth, the default, and ten times the default values for λres and λinv, and half, the default, and twice the default radius ϵ.
Results are single-run classification accuracies. The accuracy ranges are 0.6 and 0.5 percentage points for the two loss weights and 0.8 points for the perturbation radius. The default configuration performs best among the tested values, with limited degradation elsewhere. These results indicate bounded performance variation within the tested ranges, while showing that regularization strength and perturbation radius still require an appropriate balance.
The ablations in Table 2 assess whether each auxiliary objective contributes, whereas these additional analyses examine sensitivity to their weights and the perturbation radius. We will present both in the revised manuscript to provide a more complete justification of the design choices.