Semi-supervised Training Data Generation for Multilingual Question Answering

By Kyungjae Lee, Kyoungho Yoon, Sunghyun Park, Seung-won Hwang. 2018

47 citations recorded by the source corpus.

Recently, various datasets for question answering (QA) research have been released, such as SQuAD, Marco, WikiQA, MCTest, and SearchQA. However, such existing training resources for these task mostly support only English. In contrast, we study semi-automated creation of the Korean Question Answering Dataset (K-QuAD), by using automatically translated SQuAD and a QA system bootstrapped on a small QA pair set. As a na¨ıve approach for other language, using only machine-translated SQuAD shows limited performance due to translation errors. We study why such approach fails and motivate needs to build seed resources to enable leveraging such resources. Speciﬁcally, we annotate seed QA pairs of small size (4K) for Korean language, and design how such seed can be combined with translated English resources. These approach, by combining two resources, leads to 71.50 F1 on Korean QA (comparable to 77.3 F1 on SQuAD).

Subfields: Artificial Intelligence, Natural Language Processing

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

In Natural Language Processing

Citations in the corpus

MLQA: Evaluating Cross-lingual Extractive Question Answering (2019, methodology): “4 Before translating, we enclose answers in quotes, as in Lee et al. (2018).”
XTREME: A Massively Multilingual Multi-task Benchmark for Evaluating Cross-lingual Generalization (2020, methodology): “We experimented with enclosing the answer span in the English context in quotes (Lee et al., 2018; Lewis et al., 2019) but found that quotes were often dropped in translations (at different rates depending on the language).”
Learning with Limited Data for Multilingual Reading Comprehension (2019, methodology): “When compared with Base3, semisupervision (Lee et al., 2018) using 2K human labels outperforms our BiDAF model (without Refinery), on Wiki Kr set.”
KorQuAD1.0: Korean QA Dataset for Machine Reading Comprehension (2019, background): “For Korean reading comprehension, K-QuAD[5] consists of 70,000 questions that are translations of SQuADv1.”
Multilingual Extractive Reading Comprehension by Runtime Machine Translation (2018, methodology): “To alleviate the scarcity of training data in non-English languages, previous work creates a new large-scale dataset for a language of interest (He et al., 2017) or combines a medium-scale dataset in the language with an existing dataset translated from English (Lee et al., 2018).”