PulseAugur
中
实时 07:10:38

Tokenization complexity: BPE approximation proven NP-hard

一篇新的研究论文探讨了诸如字节对编码(BPE)之类的分词算法的计算复杂性。该研究证明,除非 P=NP,否则没有高效的算法可以在 $c n / ext{log } n$ 的因子内近似二元字符串上 BPE 的最优合并次数。对于可以跳过出现的合并,一个简单的算法可以达到 $O(n / ext{log log } n)$,而对于一元数据集,这两种分词问题都是强 NP 完全的。 AI

影响 理解分词的复杂性对于优化大型语言模型的效率至关重要。

排序理由 学术论文,详细介绍了关于分词算法的理论计算机科学研究。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.CL 阅读 →

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

Tokenization complexity: BPE approximation proven NP-hard

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
学术论文,详细介绍了关于分词算法的理论计算机科学研究。[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
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]

  1. arXiv cs.CL TIER_1 English(EN) · Kundan Baliga, Aiden Xie, Aditya Byju ·

    二元字符串上的成对合并分词难以近似,在一元数据集上是强NP难的

    arXiv:2611.13568v1 Announce Type: cross Abstract: Tokenisers such as Byte-Pair Encoding (BPE) shorten text by repeatedly merging an adjacent pair of symbols into a new one; fewer tokens mean less computation. We ask whether an efficient algorithm can choose $k$ merges that leave …