We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.CL

Change to browse by:

cs

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Computation and Language

Title: Improving Cross-Lingual Reading Comprehension with Self-Training

Abstract: Substantial improvements have been made in machine reading comprehension, where the machine answers questions based on a given context. Current state-of-the-art models even surpass human performance on several benchmarks. However, their abilities in the cross-lingual scenario are still to be explored. Previous works have revealed the abilities of pre-trained multilingual models for zero-shot cross-lingual reading comprehension. In this paper, we further utilized unlabeled data to improve the performance. The model is first supervised-trained on source language corpus, and then self-trained with unlabeled target language data. The experiment results showed improvements for all languages, and we also analyzed how self-training benefits cross-lingual reading comprehension in qualitative aspects.
Comments: 8 pages, 4 figures
Subjects: Computation and Language (cs.CL)
Cite as: arXiv:2105.03627 [cs.CL]
  (or arXiv:2105.03627v1 [cs.CL] for this version)

Submission history

From: Wei-Cheng Huang [view email]
[v1] Sat, 8 May 2021 08:04:30 GMT (9381kb,D)

Link back to: arXiv, form interface, contact.