我偶然发现了这些幻灯片,关于CNN使用word2vec来介绍CNN:http://www.people.fas.harvard.edu/~yoonkim/data/Kim_EMNLP_2014_slides.pdf
在幻灯片20上,第四个要点是:
Words not in word2vec are initialized randomly from U[?a, a]
where a is chosen such that the unknown words have the
same variance as words already in word2vec.
现在我想知道如何计算"a"以及如何计算完全未知单词的整个向量.