A Joint Source-Channel Model for Machine Transliteration

By Haizhou Li, Min Zhang, Jian Su. 2004

290 citations recorded by the source corpus.

Most foreign names are transliterated into Chinese, Japanese or Korean with approximate phonetic equivalents. The transliteration is usually achieved through intermediate phonemic mapping. This paper presents a new framework that allows direct orthographical mapping (DOM) between two different languages, through a joint source-channel model, also called n-gram transliteration model (TM). With the n-gram TM model, we automate the orthographic alignment process to derive the aligned transliteration units from a bilingual dictionary. The n-gram TM under the DOM framework greatly reduces system development effort and provides a quantum leap in improvement in transliteration accuracy over that of other state-of-the-art machine learning algorithms. The modeling framework is validated through several experiments for English-Chinese language pair.

Subfields: Artificial Intelligence

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

Citations in the corpus

Deep Learning: Methods and Applications (2014, background): “Many of the discriminative techniques for supervised learning in signal and information processing are shallow architectures such as HMMs [52, 127, 147, 186, 188, 290, 394, 418] and conditional random fields (CRFs) [151, 155, 281, 400, 429, 446].”
A tutorial survey of architectures, algorithms, and applications for deep learning (2014, background): “Many of the discriminative techniques in signal and information processing apply to shallow architectures such as HMMs (e.g., [87–94]) or CRFs (e.g., [95–100]).”
Jointly Optimizing a Two-Step Conditional Random Field Model for Machine Transliteration and Its Fast Decoding Algorithm (2010, methodology): “The training parallel data without alignment is ﬁrst aligned by a Viterbi version EM algorithm (Li et al., 2004).”
Transliterated Mobile Keyboard Input via Weighted Finite-State Transducers (2017, background): “Hence, much work in transliteration is focused on translation and information retrieval (Knight and Graehl, 1998; Chen et al., 1998; Virga and Khudanpur, 2003; Haizhou et al., 2004; Gupta et al., 2014).”
Hindi-to-Urdu Machine Translation through Transliteration (2010, methodology): “The character-based transliteration model pc(h|u) is computed in terms of pc(h, u), a joint character model, which is also used for ChineseEnglish back-transliteration (Li et al., 2004) and Bengali-English name transliteration (Ekbal et al.”
Processing South Asian Languages Written in the Latin Script: the Dakshina Dataset (2020, background): “…NLP work on automatic transliteration between writing systems was driven by the needs of machine translation or information retrieval systems, and hence was generally focused on proper names and/or loanwords (Knight and Graehl, 1998; Chen et al., 1998; Virga and Khudan-pur, 2003; Li et al., 2004).”
Transliteration Based Approaches to Improve Code-Switched Speech Recognition Performance (2018, unknown): “For example, in [24], transliteration between Chinese and English was achieved through a joint source channel model, referred to as an ngram-transliteration model, that aligns phonemic units be-tween two writing systems via a bilingual dictionary.”