English(EN) Robust-by-Design Distributional Learning from Contaminated Samples

新的CC-MMD方法从污染数据中学习

作者 PulseAugur 编辑部 · [1 个来源] · 2026-11-02 05:00

研究人员开发了一种名为污染校正最大均值差异（CC-MMD）的新方法，用于处理包含未知污染物的学习数据集。该方法结合了代理分数和审计校正，即使在数据严重损坏的情况下，也能准确估计潜在清洁样本的最大均值差异（MMD）。CC-MMD不需要对代理分数进行校准假设，也不需要对污染物进行结构假设，并且已被证明可以改善污染回归任务中的参数恢复和预测。该方法使在混合数据上训练的生成模型能够恢复清洁目标分布，证明了污染数据不必被丢弃。 AI

影响能够在使用真实世界中通常被污染的数据集上更鲁棒地训练机器学习模型。

排序理由详细介绍机器学习新统计方法的学术论文。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv stat.ML 阅读 →

arXiv
CC-MMD

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: 1 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 详细介绍机器学习新统计方法的学术论文。[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, other; 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: 1 days old; Coverage has settled into its steady-state source set.

完整方法见我们的编辑标准。

报道来源 [1]

arXiv stat.ML TIER_1 English(EN) · Nevena Gligi\'c, Arya Farahi · 2026-11-02 05:00

Robust-by-Design Distributional Learning from Contaminated Samples

arXiv:2610.40884v1 Announce Type: new Abstract: Modern data analysis pipelines increasingly rely on large datasets assembled through retrieval, scraping, automated filtering, and post-hoc curation, where the observed sample is often a contaminated mixture rather than the target p…

本文如何被排名

报道来源 [1]

Robust-by-Design Distributional Learning from Contaminated Samples

相关实体

相关话题