Weakly Supervised Named Entity Transliteration and Discovery from Multilingual Comparable Corpora

Proceedings of the 21st International Conference on Computational Linguistics and 44th Annual Meeting of the ACL, pages 817–824, Sydney, July 2006. c⃝2006 Association for Computational Linguistics Weakly Supervised Named Entity Transliteration and Discovery from Multilingual Comparable Corpora Alexandre Klementiev Dept. of Computer ScienceDan Roth University of Illinois Urbana, IL 61801 � klementi,danr � @uiuc.edu Abstract Named Entity recognition (NER) is an important part of many natural language processing tasks. Current approaches of- ten employ machine learning techniques and require supervised data. However, many languages lack such resources. This paper presents an (almost) unsupervised learning algorithm for automatic discov- ery of Named Entities (NEs) in a resource free language, given a bilingual corpora in which it is weakly temporally aligned with a resource rich language. NEs have similar time distributions across such corpora, and often some of the tokens in a multi-word NE are transliterated. We develop an algo- rithm that exploits both observations itera- tively. The algorithm makes use of a new, frequency based, metric for time distribu- tions and a resource free discriminative ap- proach to transliteration. Seeded with a small number of transliteration pairs, our algorithm discovers multi-word NEs, and takes advantage of a dictionary (if one ex- ists) to account for translated or partially translated NEs. We evaluate the algorithm on an English-Russian corpus, and show high level of NEs discovery in Russian. 1 Introduction Named Entity recognition has been getting much attention in NLP research in recent years, since it is seen as significant component of higher level NLP tasks such as information distillation and question answering. Most successful approaches to NER employ machine learning techniques, which require supervised training data. However, for many languages, these resources do not ex- ist. Moreover, it is often difficult to find experts in these languages both for the expensive anno- tation effort and even for language specific clues. On the other hand, comparable multilingual data (such as multilingual news streams) are becoming increasingly available (see section 4). In this work, we make two independent obser- vations about Named Entities encountered in such corpora, and use them to develop an algorithm that extracts pairs of NEs across languages. Specifi- cally, given a bilingual corpora that is weakly tem- porally aligned, and a capability to annotate the text in one of the languages with NEs, our algo- rithm identifies the corresponding NEs in the sec- ond language text, and annotates them with the ap- propriate type, as in the source text. The first observation is that NEs in one language in such corpora tend to co-occur with their coun- terparts in the other. E.g., Figure 1 shows a his- togram of the number of occurrences of the word Hussein and its Russian transliteration in our bilin- gual news corpus spanning years 2001 through late 2005. One can see several common peaks in the two histograms, largest one being around the time of the beginning of the war in Iraq. The word Russia, on the other hand, has a distinctly different temporal …

Content truncated to 500 words