Using the Web to Train a Mobile Device Oriented Japanese Input Method Editor

This paper describes the construction of a Japanese Input Method Editor (IME) system for mobile devices, using the largescale Web pages. We provide the training process of our IME model, n-pos model for local Kana-Kanji conversion and ngram model for online cloud service. Especially, we propose an online algorithm of mining new compound words, together with the detailed post-filtering process to prune the billion level entries to be used in mobile services. Experiments show that our IME system outperforms two state-ofthe-art Japanese IME baselines. We have released our system in a completely free form1 and the system is currently used by millions of users.