- Rating
- 1
- Confidence
- 3
- Informativeness
- 2
- Presentation
- 3
- Soundness
- 2
Read review
Paper summary
The paper proposes Cross-Layer Attention Recalibration (CLAR), which constructs channel and spatial gates by attending to compact representations from preceding blocks and the current block within each residual stage. Historical source dropping regularizes this retrieval. The method is evaluated on CIFAR-100, ImageNet-1K, and COCO.
Critical strengths
-
Storing channel vectors and single-channel spatial maps instead of full historical feature tensors is a sensible design. It makes individual historical sources accessible with a compact explicit memory.
-
The evaluation covers classification and dense prediction. The reported ImageNet Top-1 gains over CBAM are 1.26, 0.71, and 0.56 percentage points for ResNet-50/101/152, with small additional parameter counts over CBAM. These gains deserve recognition, particularly at depth 50, although training differences complicate attribution.
Critical weaknesses
-
Limited conceptual advance over closely related attention designs: CLAR retains CBAM’s channel-then-spatial gating order (Woo et al., 2018), and the spatial branch uses the familiar channel pooling and convolutional construction. Its principal extension is retrieval over compact historical attention representations, together with source dropping. This is a meaningful distinction from local CBAM, but historical information reuse is already central to MRLA-L (Fang et al., 2023 ICLR), DLA-L (Wang et al., 2024 IJCAI), and DIA (Huang et al., 2020 AAAI). The paper does not yet establish that this combination yields a sufficiently substantial new insight. In Table 3, gains over CBAM on CIFAR-100 are 0.21/0.40/0.67 points, and ImageNet gains over DLA-L are only 0.19/0.29/0.19 points. These numbers should be interpreted with the training-protocol caveat below.
-
The additional benefit of spatial recalibration is weakly established: In Table 1, adding spatial recalibration to the channel-only variant improves Top-1 by just 0.07 points on ResNet-50 and 0.22 on ResNet-101; ResNet-50 Top-5 decreases from 94.67 to 94.46. Spatial-only Top-1 is also below the listed ResNet references by 0.72 and 0.07 points. The stronger concern is that the matched channel-only versus full-model comparison provides limited evidence for the branch’s incremental value. No corresponding branch ablation is reported for ImageNet or COCO.
-
Limited evidence for an advantageous accuracy–cost trade-off: In Tables 4 and 5, CLAR improves the headline COCO metrics over MRLA-L by only 0.1–0.2 AP while using approximately 5.32%–7.24% more total detector parameters. The corresponding parameter increase for ImageNet classification is 9.21%–10.07%. Without matched training and uncertainty estimates, these small AP margins do not establish a robust advantage. Please report bounding-box and mask AP to two decimal places using the unrounded evaluation outputs to make the differences clearer.
-
Missing close COCO baselines: CBAM and DLA-L are omitted from the COCO tables despite their relevance to CLAR’s design. CBAM has been evaluated with Faster R-CNN on COCO (Woo et al., 2018), and DLA-L reports both Faster R-CNN and Mask R-CNN results (Wang et al., 2024). Including these methods under matched settings would clarify whether historical channel–spatial recalibration improves on the closest alternatives.
-
Relevant recent attention modules are missing from the evaluation: SE (Hu et al., 2018 CVPR), CBAM (Woo et al., 2018 ECCV), MRLA-L (Fang et al., 2023 ICLR), and DLA-L (Wang et al., 2024) remain relevant baselines; their age alone is not a weakness. However, recent work directly addresses the claimed contributions: SCSA (Si et al., 2025 Neurocomputing) studies spatial–channel interaction on ResNet backbones, including ImageNet and COCO, while the extended DIA study (Huang et al., 2025 Neucom) examines sharing and calibrating attention across layers. Positioning CLAR against these works, with a matched comparison to a suitable recent module such as SCSA, would clarify its present contribution. Achieving the latest leaderboard result is not itself a requirement.
-
Source weighting and source dropping are not disentangled: In Table 2, static and uniform weighting use no source dropping, whereas the reported dynamic configuration uses p = 0.4. The text states that dynamic weighting without dropping underperforms both alternatives, but the corresponding row is missing. Thus, the comparison does not isolate the benefit of input-dependent source weighting from regularization. Please report the missing result and compare all three weighting schemes at the same dropping probability to establish whether dynamic historical selection is necessary.
Questions for authors
-
Under one common training recipe, does CLAR improve on CBAM and DLA-L? Please prioritize the missing COCO comparisons and explain any differences in initialization, schedules, augmentation, or detector settings that affect the reported results.
-
What evidence establishes the added value of the spatial branch beyond channel-only historical recalibration? A matched branch ablation on one existing ImageNet or COCO configuration, together with per-run results where available, would help determine whether the small CIFAR gains are consistent.
-
Can you report the missing dynamic p = 0 row and compare static, uniform, and dynamic source weighting at the same dropping probability? Which gain remains attributable specifically to input-dependent historical selection?
-
What benefit does CLAR offer over DLA-L and a relevant recent module under comparable training and resource budgets? Measured memory or latency on one representative setting would help substantiate the efficiency motivation.
4 comments
(2/3) Comment for Reviewer rpVr
Contribution of the spatial branch (W2 & Q2)
We thank the reviewer for suggesting branch ablations on a larger-scale dataset. In addition to our four-GPU workstation, we rented a cloud server with the same configuration, but the matched multi-seed ImageNet comparisons and COCO baseline reproductions consumed most of our computational resources. For example, a single ImageNet ResNet-101 run takes approximately three days on two GPUs. We therefore supplemented the evaluation with single-seed branch ablations on ImageNet ResNet-50:
| Configuration | Runs | Top-1 (%) | Top-5 (%) |
|---|---|---|---|
| Spatial-only | 1 | 77.098 | 93.608 |
| Channel-only | 1 | 77.482 | 93.782 |
| Channel + spatial | 5 (mean) | 77.5592 | 93.784 |
Consistent with the CIFAR-100 ablations, the full model achieves higher mean Top-1 accuracy than either single-branch variant, with a modest gain over channel-only recalibration and little difference in ImageNet Top-5 accuracy. These observations provide preliminary evidence for the spatial branch's incremental value across datasets. We will provide these results in the appendix, explicitly label the number of runs, and note that the branch ablations have not yet been validated across multiple seeds.
Matched COCO comparisons (W3, W4 & Q1)
We thank the reviewer for pointing out the gaps in the COCO baselines and comparison settings. We added CBAM and DLA-L and reproduced all baselines and CLAR under the same detector training settings. The updated results below report AP to two decimal places. For each method, we initialized the backbone from the checkpoint of its best-performing ImageNet seed. We will detail the training schedules and data augmentation for each method in the appendix.
| Backbone | Method | Faster R-CNN box AP | Mask R-CNN box AP | Mask R-CNN mask AP |
|---|---|---|---|---|
| R50 | ResNet | 36.73 | 37.67 | 34.35 |
| R50 | SE | 38.58 | 39.27 | 35.65 |
| R50 | CBAM | 38.67 | 39.70 | 35.82 |
| R50 | MRLA-L | 40.17 | 40.81 | 36.88 |
| R50 | DLA-L | 40.08 | 41.03 | 36.94 |
| R50 | CLAR | 40.42 | 40.80 | 36.58 |
| R101 | ResNet | 39.28 | 39.96 | 36.14 |
| R101 | SE | 40.48 | 41.52 | 37.34 |
| R101 | CBAM | 40.82 | 41.59 | 37.31 |
| R101 | MRLA-L | 42.11 | 42.82 | 38.46 |
| R101 | DLA-L | 41.77 | 42.56 | 38.28 |
| R101 | CLAR | 42.42 | 43.18 | 38.39 |
Under matched settings, CLAR outperforms CBAM at both depths with both detectors. Compared with MRLA-L and DLA-L, CLAR achieves higher box AP at both depths with Faster R-CNN and with Mask R-CNN ResNet-101, but does not outperform them on Mask R-CNN ResNet-50 or all mask AP comparisons. The updated results support CLAR's detection performance while showing that the gains vary across detectors and depths. We will update the manuscript accordingly and avoid generalizing these results into a superior accuracy–cost trade-off across all configurations.
(3/3) Comment for Reviewer rpVr
Accuracy–cost comparison (W3, W5 & Q4)
We supplemented the evaluation with measured inference throughput, latency, and peak training memory. The following ImageNet ResNet-50 comparison combines five-seed Top-1 accuracy (mean ± sample standard deviation) with costs measured on an RTX 4090 in FP32. Throughput and training memory use batch size 16; latency uses batch size 1.
| Method | Top-1 (%) ↑ | Throughput (images/s) ↑ | Latency (ms) ↓ | Peak training memory (GiB) ↓ |
|---|---|---|---|---|
| CBAM | 77.07 ± 0.23 | 1643.44 | 9.39 | 2.32 |
| MRLA-L | 77.04 ± 0.10 | 1744.37 | 9.16 | 2.72 |
| DLA-L | 76.91 ± 0.10 | 824.85 | 18.84 | 3.09 |
| SCSA | 77.30 ± 0.16 | 1011.56 | 14.84 | 1.98 |
| CLAR | 77.56 ± 0.16 | 916.95 | 16.72 | 2.33 |
CLAR achieves the highest mean Top-1 accuracy among these methods. Compared with DLA-L, it also provides higher throughput, lower latency, and lower peak training memory. Compared with SCSA, the accuracy gain comes with higher runtime and memory costs. We will provide results for all evaluated depths and methods, including the detection setting, together with the detailed measurement protocol in the appendix.
Disentangling source weighting from source dropping (W6 & Q3)
We thank the reviewer for pointing out the inconsistent source-dropping settings in the original weighting comparison. We added experiments with static and uniform weighting at and included the dynamic results previously shown in Figure 2 but omitted from Table 2. The table below reports CIFAR-100 Top-1 accuracy as mean ± sample standard deviation over five random seeds.
| Weighting | ResNet-50 (%) | ResNet-101 (%) | |
|---|---|---|---|
| Uniform | 0 | 79.12 ± 0.35 | 79.27 ± 0.42 |
| Static | 0 | 78.95 ± 0.32 | 79.39 ± 0.47 |
| Dynamic | 0 | 78.70 ± 0.54 | 79.01 ± 0.10 |
| Uniform | 0.4 | 79.16 ± 0.17 | 79.45 ± 0.28 |
| Static | 0.4 | 79.11 ± 0.23 | 79.41 ± 0.28 |
| Dynamic | 0.4 | 79.25 ± 0.13 | 79.59 ± 0.27 |
All three weighting schemes follow the same dropping rule: historical sources are randomly masked, the current source is always retained, and weights are renormalized over the retained sources. Under the same source-dropping probability of , dynamic weighting achieves higher mean Top-1 accuracy than both static and uniform weighting at both depths, supporting an additional contribution from input-dependent weighting beyond source dropping alone. We will include these complete comparisons in the revised manuscript and explicitly distinguish the roles of dynamic weighting and source dropping.
Thank you again for your time and constructive feedback. We have submitted the revised manuscript incorporating the additional experiments, updated comparisons, and clarifications discussed in our response, together with the complete appendix.
We would appreciate your consideration of the revised manuscript and our responses, and would be grateful if you could reconsider your overall assessment and score in light of the additional evidence.
(1/3) Comment for Reviewer rpVr
We thank the reviewer for the valuable feedback. We have supplemented the experiments as extensively as our time and computational resources permit. We apologize for posting our response on the final day of the response period due to the substantial experimental workload. We are still refining the revised manuscript and welcome further feedback, which we will do our best to address in the revision.
Novelty comparison with related methods (W1 & W5)
We thank the reviewer for highlighting these related methods, particularly SCSA and the extended DIA study, which we overlooked in our literature review. Including these recent works will make our discussion more current and comprehensive. The core design of CLAR is to reduce the cost of historical representations while retaining individual historical sources, allowing the current layer to access each source separately and recompute its weight based on the current input. This distinguishes CLAR from related attention-based methods we discussed below.
Specifically, DIA uses an LSTM to accumulate historical information into hidden and cell states. The extended DIA study further investigates sharing attention modules across layers and using an LSTM to calibrate correlated attention maps. MRLA-L and DLA-L both propagate historical information by recursively updating the aggregated output. These methods effectively reuse historical information, but the historical sources have been summarized into recurrent states, and the current layer no longer computes a separate matching weight for each previous layer.
SCSA further explores spatial–channel synergy, using spatial attention to guide subsequent channel attention. These interactions operate within the current feature, and no historical sources are explicitly accessed.
On the other hand, MRLA-Base and DLA-Base retain explicit source interactions at higher resource costs. Within the ResNet family, the ImageNet and COCO results reported for MRLA-Base focus on ResNet-50, and its appendix reports an out-of-memory issue for ResNet-101 at the original batch size; DLA-Base only reports results on CIFAR. In our reproductions, both Base methods with ResNet-101 exhibited training divergence within the first ten epochs. These observations motivated us to explore a more scalable design for explicit historical selection.
Inspired by CBAM, CLAR stores compact channel and spatial representations for each historical source, which the current layer's queries individually matches and reweights. This reduces explicit history storage from O(LCHW) to O(L(C+HW)) while preserving source-specific retrieval. Our ResNet-152 experiments demonstrate the feasibility and effectiveness of this design in deep networks.
The table below compares these methods across key design dimensions, highlighting the novel aspects of CLAR relative to prior studies.