- Rating
- 1
- Confidence
- 3
- Informativeness
- 2
- Presentation
- 2
- Soundness
- 1
Read review
Paper summary
This paper studies test-time OOD detection with vision-language models under ID-dominated streams and attributes cache corruption to a mismatch between confidence-based admission and hard-negative retrieval. It proposes a training-free mutual-kNN visual-text graph with multi-scale disparity scores and a topological cache gate, reporting strong Near-OOD results across imbalance ratios and backbones but weaker ImageNet Far-OOD performance.
Critical strengths
- The studied problem is important.
- The reported Near-OOD gains are consistent across imbalance ratios and backbones.
- The evaluation includes useful ablations and semantic-alignment visualizations.
Critical weaknesses
- The studied problem is important.
- The reported Near-OOD gains are consistent across imbalance ratios and backbones.
- The evaluation includes useful ablations and semantic-alignment visualizations.
Questions for authors
-
The realistic-streaming motivation is not established. Section 4.2 creates imbalance by subsampling OOD datasets, but does not specify stream order, bursts, batch composition, cache resets, update timing, or full-spectrum mixing. Please provide the complete online protocol and justify how it represents deployment streams.
-
The claimed cause of cache failure is not isolated. and the comparison do not form a complete negative-label-selector × cache-updater control. Please report this factorial ablation and the true ID/OOD/boundary composition of both caches over time.
-
The cache gate may be undefined or degenerate for small batches. Eqs. (8)–(9) use on a current batch of size , but is unreported; when , visual Top- is undefined or contains all samples, and for the skeleton connects every text to the same image. What was used, and how stable is the gate across , , and ?
-
The theoretical result does not derive the algorithm. Eqs. (3)–(4) provide an algebraic analogy to message passing, while Appendix A.1 assumes the exterior-shell model and the similarity orderings required by Eqs. (22)–(30). Which assumptions are verified, and what result specifically supports mutual-NN, the skeleton, and Eq. (18)?
-
The necessity of several components remains unclear. Table 2 ablates disparity scores and , but not mutual-NN versus the Top-1 skeleton, text-based KMeans initialization, or a simple cosine/NN cache gate. Please include these controls to distinguish the proposed mechanism from a simpler geometric rule.
-
The conclusions should reflect the observed Near-/Far-OOD trade-off. Table 3 shows lower ImageNet Far-OOD AUROC and higher FPR@95 than TANL for every backbone, whereas Table 4 shows the opposite trend on CIFAR-100. Please explain this inconsistency and scope the robustness claim accordingly.
-
The statistical evidence is difficult to interpret. Table 1 averages over five different OOD ratios, and several Figure 2 margin error bars cross zero despite the claim of “strictly positive and wide” margins. Please provide per-ratio numerical results, the source of randomness, confidence intervals or significance tests, and a more qualified mechanism claim.
-
The computational and reproducibility details are incomplete. The paper omits , graph-recomputation frequency, , runtime, peak memory, latency, the negative-text corpus, prompt template, KMeans details, and stream seeds. Please report these items and compare efficiency with TANL and the other test-time baselines.