基于无标签非一致性得分的半监督保形预测
Semi-Supervised Conformal Prediction With Unlabeled Nonconformity Score
为视觉模型提供一种在标注数据稀缺时仍能保证覆盖率的校准方法，可直接用于分类、分割等任务的不确定性量化，提升预测集在不同运行间的稳定性。
原文依据 we extend CP to the semi-supervised setting and propose SemiCP, a new paradigm that leverages both labeled and unlabeled data for calibration.
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论保形预测（CP）是一种强大的不确定性量化框架，能够生成具有覆盖率保证的预测集。分裂保形预测在校准过程中依赖有标签数据。然而，现实场景中有标签数据往往有限，导致不同运行间的覆盖率性能不稳定。为解决该问题，我们将CP扩展到半监督设定，并提出SemiCP，一种利用有标签和无标签数据进行校准的新范式。为此，我们引入一种无标签非一致性得分——最近邻匹配（NNM）得分。具体而言，NNM在校准过程中利用无标签样本最相似的伪标签对应样本来估计其非一致性得分，同时保持有标签数据的原始得分不变。理论上，我们证明SemiCP的平均覆盖率差距（即经验边际覆盖率与目标覆盖率之间的绝对差）能够以O(N^{-1}t^2)的速率显著下降并收敛到一个误差项，其中N为无标签数据的数量。大量实验验证了SemiCP在有标签数据有限时的有效性：在仅有20个有标签样本的情况下，使用4000个无标签样本在常见基准上可将平均覆盖率差距最多降低77%。代码见 https://github.com/Shinning-Zhou/SemiCP。
英文原文摘要 1386 字符
Conformal prediction (CP) is a powerful framework for uncertainty quantification, generating prediction sets with coverage guarantees. Split conformal prediction relies on labeled data in the calibration procedure. However, the labeled data is often limited in real-world scenarios, leading to unstable coverage performance in different runs. To address this issue, we extend CP to the semi-supervised setting and propose SemiCP, a new paradigm that leverages both labeled and unlabeled data for calibration. To achieve this, we introduce an unlabeled nonconformity score, Nearest Neighbor Matching (NNM) score. Specifically, NNM estimates the nonconformity scores of unlabeled samples using their most similar pseudo-labeled counterparts during calibration, while maintaining the original scores for labeled data. Theoretically, we demonstrate that the average coverage gap (i.e., the absolute difference between the empirical marginal coverage and the target coverage) of SemiCP can decrease significantly at a rate O(N-1t2) and converge to an error term, where N is the number of unlabeled data. Extensive experiments validate the effectiveness of SemiCP under limited labeled data, reducing the average coverage gap by up to 77% on common benchmarks with 4000 unlabeled examples, when there are only 20 labeled examples. Code is available at https://github.com/Shinning-Zhou/SemiCP.