Apr 25, 2018 · Artificial Intelligence

How cw2vec Beats Word2Vec: Leveraging Chinese Stroke N‑grams for Superior Word Embeddings

This article introduces cw2vec, a novel Chinese word‑embedding algorithm that exploits stroke‑level subword information, outlines its theoretical foundations, compares it with word2vec, GloVe, CWE and other models on multiple benchmarks, and demonstrates its superior performance across word similarity, analogy, text classification and named‑entity recognition tasks.

Chinese NLPUnsupervised Learningcw2vec

0 likes · 14 min read

How cw2vec Beats Word2Vec: Leveraging Chinese Stroke N‑grams for Superior Word Embeddings

AntTech

Jan 18, 2018 · Artificial Intelligence

cw2vec: Learning Chinese Word Embeddings with Stroke n-grams

The cw2vec paper, presented at AAAI 2018, introduces a Chinese word embedding method that leverages stroke n‑grams to capture character semantics, proposes a novel loss function, demonstrates consistent improvements over existing models across similarity, analogy, classification and NER tasks, and discusses real‑world AI applications.

AAAI 2018AI researchChinese NLP

0 likes · 7 min read

cw2vec: Learning Chinese Word Embeddings with Stroke n-grams