超越平面标签：面向层次细粒度视觉分类的层级限制对比学习
Beyond Flat Labels: Level-Restricted Contrastive Learning for Hierarchical Fine-Grained Vision Classification
AI总结：
针对多模态对比学习在层次标签空间中预测不一致的问题，提出层级限制对比学习，仅在同一层级内进行对比，并采用组平衡设计，显著提升层次一致性和分类精度。
摘要：
多模态对比学习通过将图像与文本类别对齐，实现了零样本视觉分类。然而，在层次结构化的标签空间中，现有方法往往产生跨分类层级不一致的预测。例如，模型可能预测一个细粒度类别，但其父类别与同时预测的高层级标签相矛盾。通过分析，问题源于对比比较涉及多个分类层级时产生的假负标签。为此，我们提出将对比比较限制在同一分类层级内的类别之间。此外，我们采用组平衡设计，确保每个分类层级获得充分的优化。结果，所提框架从粗粒度到细粒度提高了层次一致性和分类准确性。我们基于BioCLIP使用TreeOfLife-10M训练模型，并在多个层次分类基准上进行评估，模型在欧几里得空间和双曲空间中都表现出显著改善的层次一致性。值得注意的是，在iNaturalist 2021（iNat21）上，我们的方法在各层级的平均准确率比基线提高了30.47%，突显了其在层次零样本分类中的有效性。
英文摘要：
Multimodal contrastive learning has enabled zero-shot visual classification by aligning images with textual categories. However, in hierarchically structured label spaces, existing methods often produce predictions that are inconsistent across taxonomic levels. For example, a model may predict a fine-grained category whose parent category contradicts its simultaneously predicted higher-level label. By analysis, the issue originates from false negative labels when contrastive comparison involves multiple taxonomic levels. To this end, we propose to restrict contrastive comparisons to categories within the same taxonomic level. In addition, we adopt a group-balanced design, ensuring each taxonomic level receives adequate optimization. As a result, the proposed framework improves both hierarchical consistency and classification accuracy from coarse to fine granularity. We train our model with TreeOfLife-10M based on BioCLIP and evaluate it across multiple hierarchical classification benchmarks, where the model demonstrates significantly improved hierarchical consistency in both Euclidean and hyperbolic spaces. Notably, on iNaturalist 2021 (iNat21), our method improves average accuracy across levels by 30.47% over the baseline, highlighting its effectiveness for hierarchical zero-shot classification.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
