CoFailMed：医学分割冻结编码器模型池的共失败审计
Different Models, Shared Failures: Auditing Frozen-Encoder Pools for Medical Segmentation
AI 导读
CoFailMed 对医学分割预训练编码器模型池的共失败情况进行了审计。研究在 5 个 2D 任务上评估 11 个公开预训练编码器（每编码器 4 个随机种子），发现共享 checkpoint 的模型对平均逐例 Dice 相关性最高；17 编码器扩展分析显示同家族 checkpoint 间的相关性仍高于跨家族组合。
正文 · 原文
Abstract:Average segmentation accuracy and model count do not reveal whether models fail on the same cases. We present CoFailMed, an audit of correlated failures in pretrained-encoder pools for medical segmentation. Across five 2D tasks, we evaluate 11 public pretrained encoders (a controlled sample, not a census) using frozen features and separately trained $1\times 1$ readouts with four seeds per encoder. Shared-checkpoint pairs have the highest mean per-case Dice correlation. An expanded 17-encoder analysis shows that distinct checkpoints within the same encoder family remain more correlated than cross-family pairs. At the pooled 33rd-percentile Dice threshold, four-member pools drawn from distinct encoders have mean effective sizes of 1.56-2.92, estimated from failure correlations in pools with enough failures and passes per member. In the primary pool, richer readouts and fine-tuning narrow the Dice-correlation gap between pairs sharing a checkpoint (or the DINOv2-B/S release) and all other pairs. Yet dependence between different encoders can increase, and fine-tuning does not consistently increase effective pool size at 33rd-percentile thresholds recomputed for each training recipe. A smaller between-group gap can coexist with greater dependence. Pool evaluation should therefore report absolute dependence and threshold-specific effective size alongside accuracy and nominal pool size. The CoFailMed release provides aligned per-case traces, analysis scripts and study code to reproduce the central audit results.
| Comments: | Accepted at NeurIPS 2026, Track on Evaluations and Datasets. 47 pages, 8 figures. Code and data: this https URL |
| Subjects: | Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.41790 [cs.CV] |
| (or arXiv:2610.41790v1 [cs.CV] for this version) | |
| https://doi.org/10.48550/arXiv.2610.41790 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Eunseob Choi [view email]
[v1]
Fri, 30 Oct 2026 14:21:52 UTC (1,326 KB)
来源：arXiv cs.AI · arxiv.org