Keyword translation accuracy and cross-lingual question answering in Chinese and Japanese

Mitamura, Teruko; Wang, Mengqiu; Shima, Hideki; Lin, Frank

doi:10.3115/1708097.1708104

Cited by 6 publications

(4 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this case, because of the lack of the crucial relevant information, the MT system cannot distinguish between the correct translations of "Mercury" and "first-person" from other words with the same-spelling. This kind of issue in MT has been studied by previous work in MLQA (Mitamura et al, 2006;Ture and Boschee, 2016). In addition, in extractive RC, the answer spans are subphrases in the context paragraph, so generating answer spans by back-translation is not a desirable approach, as it generates homographic variations.…”

Section: Drawback Of the Back-translation Systemmentioning

confidence: 99%

See 1 more Smart Citation

Multilingual Extractive Reading Comprehension by Runtime Machine Translation

Asai¹,

Eriguchi²,

Hashimoto³

et al. 2018

Preprint

View full text Add to dashboard Cite

Despite recent work in Reading Comprehension (RC), progress has been mostly limited to English due to the lack of large-scale datasets in other languages. In this work, we introduce the first RC system for languages without RC training data. Given a target language without RC training data and a pivot language with RC training data (e.g. English), our method leverages existing RC resources in the pivot language by combining a competitive RC model in the pivot language with an attentive Neural Machine Translation (NMT) model. We first translate the data from the target to the pivot language, and then obtain an answer using the RC model in the pivot language. Finally, we recover the corresponding answer in the original language using soft-alignment attention scores from the NMT model. We create evaluation sets of RC data in two non-English languages, namely Japanese and French, to evaluate our method. Experimental results on these datasets show that our method significantly outperforms a back-translation baseline of a state-of-the-art product-level machine translation system.

show abstract

Section: Drawback Of the Back-translation Systemmentioning

confidence: 99%

“…It is a common approach in MLQA to translate all non-English text or keyword in queries into English beforehand, and then to treat the task as a monolingual task (Ture and Boschee, 2016;Mitamura et al, 2006;Hartrumpf et al, 2009;Esplà-Gomis et al, 2012).…”

Section: Related Workmentioning

confidence: 99%

Multilingual Extractive Reading Comprehension by Runtime Machine Translation

Asai¹,

Eriguchi²,

Hashimoto³

et al. 2018

Preprint

View full text Add to dashboard Cite

show abstract

“…Ture and Lin described three methods for translating queries into the collection language in a probabilistic manner, improving document retrieval effectiveness over a one-best translation approach (2014). Extending this idea to MLQA appears as a logical next step, yet most prior work relies solely on the one-best translation of questions or answers (Ko et al, 2010b;García-Cumbreras et al, 2012;Chaturvedi et al, 2014), or selects the best translation out of few options (Sacaleanu et al, 2008;Mitamura et al, 2006). Mehdad et al reported improvements by including the top ten translations (instead of the single best) and computing a distance-based entailment score with each (2010).…”

Section: Related Workmentioning

confidence: 99%

Learning to Translate for Multilingual Question Answering

Ture¹,

Boschee²

2016

Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing

View full text Add to dashboard Cite

In multilingual question answering, either the question needs to be translated into the document language, or vice versa. In addition to direction, there are multiple methods to perform the translation, four of which we explore in this paper: word-based, 10-best, contextbased, and grammar-based. We build a feature for each combination of translation direction and method, and train a model that learns optimal feature weights. On a large forum dataset consisting of posts in English, Arabic, and Chinese, our novel learn-to-translate approach was more effective than a strong baseline (p < 0.05): translating all text into English, then training a classifier based only on English (original or translated) text.

show abstract

“…But in different systems two or adjoining modules may be realized in one module. For example, in JAVELIN question [3,4] analysis and query building are realized in one module. In this paper, AnswerOnWeb realizes answer extraction and answer selection in one module.…”

Section: Related Workmentioning

confidence: 99%

Applied-Information Technology in Intelligent Question-Answering System Based on Knowledge

Zhang

Zheng

2014

AMR

View full text Add to dashboard Cite

The AnswerOnWeb is intelligent question-answering system which network-oriented special field knowledge. On this Paper two technology of knowledge mining and knowledge annotator are used. The former is used in mining the answer which is hidden in semi-structure documents returned by internet search engine. The latter is used in trying to locate the answer in the some knowledge documents of special website or network service. The experiments proved that this system can not only get higher answer precision but also better question recall when is applied in the special field knowledge on net.

show abstract

Keyword translation accuracy and cross-lingual question answering in Chinese and Japanese

Cited by 6 publications

References 6 publications

Multilingual Extractive Reading Comprehension by Runtime Machine Translation

Multilingual Extractive Reading Comprehension by Runtime Machine Translation

Learning to Translate for Multilingual Question Answering

Applied-Information Technology in Intelligent Question-Answering System Based on Knowledge

Contact Info

Product

Resources

About