Back transliteration from Japanese to English using target English context

Abstract

This paper proposes a method of automatic back transliteration of proper nouns, in which a Japanese transliterated-word is restored to the original English word. The English words are created from a sequence of letters; thus our method can create new English words that are not registered in dictionaries or English word lists. When a katakana character is converted into English letters, there are various candidates of alphabetic characters. To ensure adequate conversion, the proposed method uses a target English context to calculate the probability of an English character or string corresponding to a Japanese katakana character or string. We confirmed the effectiveness of using the target English context by an experiment of personal-name back transliteration.

Machine transliteration and transliterated text retrieval: a survey

“…Quite a number of transliteration mechanisms have been proposed for some non-English European languages, Russian [6][7][8] and East Asian languages like Chinese [9][10][11][12], Japanese [13][14][15][16][17], Korean [18][19][20][21][22], West Asian languages like Arabic [23][24][25] and the Persian [26,27]. There have been some recent attempts on some Indian languages like Hindi [8,[28][29][30][31][32][33][34][35][36][37][38][39], Bengali [33,[40][41][42], Punjabi [43], Telugu [44], Kannada [29,45,46] and Tamil [29,31,47].…”

“…There have been a number of works using transliteration generation. Transliteration models have been proposed for the transformation mainly between English and non-English languages, including Arabic [23,[53][54][55], Persian [26], Korean [20,21,56,57], Chinese [11,[58][59][60][61][62][63], Japanese [15,[64][65][66][67][68] and Roman languages [69][70][71][72], which are discussed in the seminal survey by Karimi et al [48].…”

Fast decoding and easy implementation

“…Most previous transliteration studies have relied on a generative model resembling the IBM model (Brown et al, 1993). This approach is applicable to various languages: for Japanese (Goto et al, 2004;Knight and Graehl, 1998), Korean (Oh and Choi, 2002;Oh and Choi, 2005;Oh and Isahara, 2007), Arabic(Stalls and Sherif and Kondrak, 2007), Chinese (Li et al, 2007), and Persian (Karimi et al, 2007). As described previously, the proposed discriminative approach differs from them.…”