BERT Based Multilingual Machine Comprehension in English and Hindi

2 Jun 2020Somil GuptaNilesh Khade

Multilingual Machine Comprehension (MMC) is a Question-Answering (QA) sub-task that involves quoting the answer for a question from a given snippet, where the question and the snippet can be in different languages. Recently released multilingual variant of BERT (m-BERT), pre-trained with 104 languages, has performed well in both zero-shot and fine-tuned settings for multilingual tasks; however, it has not been used for English-Hindi MMC yet... (read more)

PDF Abstract

Results from the Paper


TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK USES EXTRA
TRAINING DATA
RESULT BENCHMARK
Multilingual Machine Comprehension in English Hindi Extended XQuAD m-BERT augmented with Hindi QA F1 (QE-PE) 76.51 # 1
F1 (QE-PH) 57.31 # 1
F1(QH-PE) 51.04 # 1
F1(QH-PH) 59.80 # 1
EM(QE-PE) 64.29 # 1
EM(QE-PH) 44.71 # 1
EM(QH-PE) 41.01 # 1
EM(QH-PH) 45.63 # 1

Methods used in the Paper