English(EN) Metric-Dependent Annotation Saturation for Learning from Label Distributions

Apple ML Research：标注需求因评估指标而异

作者 PulseAugur 编辑部 · [1 个来源] · 2026-06-23 00:00

Apple Machine Learning Research 发表了一篇论文，详细介绍了一种名为“Metric-Dependent Annotation Saturation”（依赖于指标的标注饱和度）的方法。该方法提出，从标签分布中捕获有意义信号所需的标注者数量取决于所使用的具体评估指标。例如，在自然语言推理（NLI）模型中，实现熵相关性收敛比实现分布匹配需要更多的标注者。研究还强调，软标签（代表细微的决策边界）比独热标签提供更好的正则化和泛化能力，尤其是在处理有噪声的标注时。 AI

影响建议根据评估指标优化标注预算，以改进模型训练。

排序理由 Apple ML Research部门发表的研究论文。[lever_c_demoted from research: ic=1 ai=1.0]

在 Apple Machine Learning Research 阅读 →

AI 生成摘要 · Google Gemini · 来自 1 个来源。我们如何撰写摘要 →

本文如何被排名

Signal score: 0 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; Apple ML Research部门发表的研究论文。[lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration: Single-source cluster; Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics: paper, model release; Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance: High; Clearly on-topic for AI-industry coverage.
Story freshness: 99 days old; Aged out of breaking-news scoring windows; ranking reflects the durable signal from the full source set.

完整方法见我们的编辑标准。

报道来源 [1]

Apple Machine Learning Research TIER_1 English(EN) · 2026-06-23 00:00

面向标签分布学习的度量依赖性标注饱和度

When annotators disagree on a label, the disagreement itself carries signal—and the number of annotators needed to capture it depends on the evaluation metric. We fine-tune NLI models on label distributions subsampled from ChaosNLI, a dataset providing 100 independent annotator j…

本文如何被排名

报道来源 [1]

面向标签分布学习的度量依赖性标注饱和度

相关实体

相关话题