- Rating
- 2
- Confidence
- 3
- Informativeness
- 2
- Presentation
- 2
- Soundness
- 2(+1)
Read review
Paper summary
This paper studies the numerical instability of Poincaré-ball representations near the boundary and proposes a truncated Poincaré quasi-metric. It retains standard Poincaré geometry within radius , switches to scaled Euclidean geometry outside, and uses a Ray Intersection Approximation (RIA) to approximate cross-boundary distances. The paper further analyzes near-boundary gradient stability and evaluates the method on 12 datasets spanning image, text, video, and cross-modal retrieval. The overall goal is to preserve the advantages of Poincaré representations while mitigating boundary-induced numerical instability.
Critical strengths
- The method is simple and directly addresses a practical problem. Near-boundary numerical instability is an important issue for Poincaré representations. By truncating the outer region and using a closed-form RIA, the method provides a direct stabilization mechanism with low additional computational complexity and can be readily integrated into existing retrieval pipelines.
- The experimental coverage is broad. The paper evaluates image, text, video, and cross-modal retrieval across 12 datasets, and further includes analyses of hyperparameter sensitivity, model scale, and embedding-norm dynamics.
Critical weaknesses
- The proof of Theorem 1 is inconsistent with the RIA actually defined in the paper. In the main text and Appendix B, the cross-boundary intersection point is defined as , i.e., the intersection of the line segment from to with the truncation sphere. However, Appendix C instead uses when proving the general- case, which is a radial projection from the origin. Thus, the current proof analyzes a different construction from the general RIA defined in the method. In addition, Eq. (25) only establishes a constant upper bound on the gradient, which directly supports rather than the claimed . This issue directly affects the paper’s main theoretical conclusion.
- The experimental setup contains inconsistencies that weaken the reliability of the performance and stability conclusions. For image retrieval, the main text states that LoRA is applied to all transformer blocks, whereas Appendix E.1 states that only the last two transformer blocks are fine-tuned, with different epoch and batch-size settings. For text retrieval, the baseline compare-all runs and the tuned truncated-Poincaré runs use different batch sizes; for InfoNCE, this changes the number of in-batch negatives. In addition, the definitions and measurement protocols of the core stability metrics in Figure 3 are not sufficiently clear. These issues make it difficult to determine whether the reported gains arise from the proposed geometry or from differences in training and tuning.
Questions for authors
- Regarding Theorem 1, could the authors provide a complete gradient-boundedness proof for the RIA actually used in the paper, ? Please also clarify why the conclusion should be rather than only .
- Regarding experimental fairness and reproducibility, please specify the exact training configurations corresponding to the main-table results and re-evaluate the major baselines under the same batch size, optimizer, training schedule, and tuning budget. Please also provide precise mathematical definitions and measurement protocols for MeanRelErr, MaxRelErr, Max distortion, and Diameter delta in Figure 3.