English(EN) CP-MLPs: A Tensor-Rank Theory of Tied and Untied MLP Blocks

新的张量框架分析Transformer MLP设计选择

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

研究人员引入了CP-MLPs，一个新颖的张量框架，用于分析Transformer前馈块内的设计选择。该框架区分了乘法单元中绑定和非绑定的参数共享，揭示了对于某些类型的交互，非绑定块提供了更高的宽度效率。对训练过的门控MLP进行的实验表明，在考虑匹配的输入投影预算时，非绑定块保持优势，尽管具体优势取决于目标结构。 AI

影响这项研究为理解和潜在优化Transformer前馈网络的设计提供了一个理论视角。

排序理由该项目是一篇学术论文，详细介绍了用于分析神经网络架构的新理论框架和实验结果。[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 →

本文如何被排名

Signal score: 34 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 该项目是一篇学术论文，详细介绍了用于分析神经网络架构的新理论框架和实验结果。[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: Breaking (< 6h); Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

完整方法见我们的编辑标准。

报道来源 [1]

arXiv cs.LG TIER_1 English(EN) · Md Rifat Arefin, Farzaneh Heidari, Irina Rish, Guillaume Rabusseau · 2026-11-11 05:00

CP-MLPs：一种张量秩理论，用于绑定和非绑定的MLP块

arXiv:2611.13647v1 Announce Type: new Abstract: Transformer feed-forward blocks combine several design choices: squared activations, gating, and tied versus untied parameter sharing, whose separate contributions to expressivity remain poorly understood. We introduce CP-MLPs, a te…

本文如何被排名

报道来源 [1]

CP-MLPs：一种张量秩理论，用于绑定和非绑定的MLP块

相关实体

相关话题