PulseAugur
中
实时 11:17:04
English(EN) Flat Loss, Evolving Features: Grokking in Modular Addition

研究发现:神经网络可在训练损失平坦的情况下学习

研究人员证明,即使在训练损失相对平坦的情况下,神经网络也能继续学习并提高性能。他们使用ReLU网络和特定的更新规则来处理模块化加法任务,观察到隐藏特征在演化,而保留的准确性显著提高,尽管训练交叉熵保持稳定。这种被称为“领悟”(grokking)的现象表明,最小化交叉熵并非正确分类的必要条件,因为预测置信度也会影响该指标。研究进一步显示,在更深的网络和Transformer模型中也存在平坦损失与学习并存的现象,这表明训练损失的平台期不一定意味着学习的终结。 AI

影响 挑战了最小化训练损失是模型改进唯一驱动力的假设,为训练更鲁棒的模型开辟了新途径。

排序理由 该集群包含一篇详细介绍机器学习新发现的学术论文。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.LG 阅读 →

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

研究发现:神经网络可在训练损失平坦的情况下学习

本文如何被排名

Signal score
9 / 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机器学习新发现的学术论文。[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, 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]

  1. arXiv cs.LG TIER_1 English(EN) · Akash Kumar, Karthik Sankaranarayanan ·

    平坦损失,演进特征:模块化加法中的 Grokking

    arXiv:2610.22593v1 Announce Type: new Abstract: Can a network keep learning while its training cross-entropy stays nearly flat at a positive level? We show that it can. On modular addition, a standard testbed for grokking, ReLU networks trained with supported-Polar hidden updates…