English(EN) OptiSelect: How does the Optimizer Shape Data Curriculum?

新的OptiSelect框架优化了LLM预训练的数据选择

作者 PulseAugur 编辑部 · [1 个来源] · 2026-10-05 04:00

研究人员推出OptiSelect，一个用于优化大型语言模型预训练期间数据选择的新框架。该框架将优化器感知选择的概念形式化，该概念考虑了优化器步骤如何影响数据候选者的价值。理论分析表明，与Lion和Muon等基于符号或极性切线的预条件器相比，AdamW和Sophia等对角自适应优化器在此选择过程中更有效。使用124M和720M模型的实验结果支持了这些发现，表明即使使用Muon作为优化器，AdamW的评分几何形状也表现最佳。 AI

影响为LLM预训练中优化器和数据选择的协同设计提供了理论基础和实践指导。

排序理由该集群包含一篇详细介绍用于优化LLM预训练的新框架和理论分析的研究论文。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.AI 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: 11 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 该集群包含一篇详细介绍用于优化LLM预训练的新框架和理论分析的研究论文。[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: Same-day; Cluster formed today. Ranking reflects the current source set at time of score.

完整方法见我们的编辑标准。

报道来源 [1]

arXiv cs.AI TIER_1 English(EN) · Simin Fan, Alireza Abdollahpoorrostam, Martin Jaggi · 2026-10-05 04:00

OptiSelect：优化器如何塑造数据课程？

arXiv:2610.03432v1 Announce Type: cross Abstract: Online data selection has demonstrated substantial efficiency gains for LLM pretraining by training on the most valuable candidates within each batch. Since a candidate's value is realized through its effective model update, princ…

本文如何被排名

报道来源 [1]

OptiSelect：优化器如何塑造数据课程？

相关实体

相关话题