按年被提到
20132015 · 4 次2026
大模型与语言模型›词向量与预训练模型·arXiv:1301.3781 · cs.CL

Efficient Estimation of Word Representations in Vector Space

向量空间中词表示的高效估计
Tomas Mikolov, Kai Chen, Greg Corrado, Jeffrey Dean · 2013

提出 word2vec,几乎不费算力就能学出高质量词向量

开始阅读 ↗PDFalphaXiv 讨论
被推荐 · 14 人首次 2013-08-15 · 最近 2025-05-29
14不同的人
14活跃月份
15讨论串
25,992被引(OpenAlex)

论文提出两种新的模型架构,用于从非常大的数据集中计算词的连续向量表示。在词相似度任务上,它们以低得多的计算成本取得大幅精度提升:在 16 亿词的数据集上不到一天就能学到高质量词向量,并在衡量句法和语义相似度的测试集上达到当时最好的表现。

HN 为什么提它

HN 上的人把它当作词向量和 embedding 的起点来推荐,称无需 NLP 或神经网络背景也读得懂,并常以「king 减 man 加 woman」的向量类比来说明它的影响。

奠基之作难度 3 / 5约 40 分钟经得起时间
读之前最好知道
1神经网络与词的向量表示
2语言模型基础(n-gram 与神经语言模型)
3softmax 与梯度下降
4词相似度与类比评测的概念

HN 书评

推荐在前,同类按时间 · 每篇最多保留 20 条,统计用全部提及
推荐magicalist · 2013-08

“but the paper is still a good read), especially [1]”

认为这篇仍值得一读,尤其讲 word2vec 把词映射为向量
word2vec – Tool for computing continuous distributed representations of words原评论 ↗
推荐3JPLW · 2013-09

“And the paper itelf is a very worthwhile read: Efficient Estimation of Word Representations in Vector Space”

推荐论文本身,值得一读,讲 word2vec 用向量表示词
word2vec in yhat: Word vector similarity原评论 ↗
推荐dhammack · 2013-11

“The paper associated with word2vec is fairly easy to read and understand, even if you don't have a background in NLP or neural networks: Efficient Estimation of Word Representations in Vector Space”

推荐这篇 word2vec 论文,没有 NLP 或神经网络背景也读得懂
Show HN: A search engine that lets you 'add' words as vectors原评论 ↗
推荐nl · 2015-06

“Word2Vec kind of exploded everyone's mind for a while.”

把它列入按时间顺序的阅读清单,称 word2vec 曾震撼众人
Hound voice search NLP demo [video]原评论 ↗
推荐danieldk · 2015-06

“Mikolov, et al. 2013 [1] do a proper evaluation of this.”

指出 Mikolov 等做了严谨评估,skip-gram 类比准确率约 50% 和 56%
Abstractions my Deep Learning word2vec model made原评论 ↗
推荐sp332 · 2015-06

“Word2vec is exciting because it has a useable amount of accuracy for a relatively small amount of cleverness on the part of the human.”

称 word2vec 几乎不靠人工巧思就有可用的准确率,令人兴奋
Abstractions my Deep Learning word2vec model made原评论 ↗
看全部 16 条 →
推荐PeterisP · 2023-10

“the key two approaches to read up on would probably be Mikolov's word2vec”

把 Mikolov 的 word2vec 列为必读的两大方法之一
Embeddings: What they are and why they matter原评论 ↗
引用deepnet · 2015-12

“The recent emergent semantic vector algebra discovered in human languages by Mikolov's word2vec”

拿 word2vec 说明词向量能做类比运算,如 king 减 man 加 woman
John Searle: Consciousness in Artificial Intelligence [video]原评论 ↗
引用deepnet · 2016-05

“Mikolov's[3] Word2vec & Radford's Image Vectors[4] produce semantic spaces where vector math is akin to reasoning.”

拿 word2vec 作例子,说明向量运算类似推理的语义空间
My path to OpenAI原评论 ↗
引用deepnet · 2017-07

“The other widely reported vector algebras in a semantic space were discovered by Mikolov et al when producing ~300 dimensional vectors for a billion word Wikipedia corpus.”

提到 Mikolov 等在维基语料上发现的词向量代数
When not to use deep learning原评论 ↗
引用nl · 2017-08

“if you have some kind of embedding of your words/tokens/sentences/something you cite Efficient Estimation of Word Representations in Vector Space (Word2Vec, Mikolov).”

指出使用词嵌入时应引用 Word2Vec 这篇
Do I really have to cite an arXiv paper?原评论 ↗
引用jventura · 2017-09

“what I could grasp from the wikipedia article [0] and the original paper [1] is that word2vec uses a vector to represent words in a n-dimensional space.”

读完维基和原论文后理解 word2vec 是在 n 维空间用向量表示词
What every software engineer should know about search原评论 ↗
引用ninjin · 2019-02

“akin to how Mikolov et al. reconsidered how to learn vector representations back in 2013 [3]”

拿 Mikolov 等 2013 年重新思考向量表示学习作类比
Better Language Models and Their Implications原评论 ↗
引用Olreich · 2023-10

“Here’s a paper about the technique which might help: Efficient Estimation of Word Representations in Vector Space”

贴出原论文,说明 word2vec 技术可帮到对方
Embeddings: What they are and why they matter原评论 ↗
引用HPsquared · 2025-02

“And the original word2vec paper by Mikolov mentions it Efficient Estimation of Word Representations in Vector Space”

指出 word2vec 原论文提到过 king-queen 类比
Why Does Integer Addition Approximate Float Multiplication?原评论 ↗
引用godelski · 2025-05

“a bit of discussion about this in the Word2Vec paper[1]”

指出 Word2Vec 论文里讨论过词嵌入是否捕捉语义相似性
A visual exploration of vector embeddings原评论 ↗
被提交为 story · 1 次