2021.03.10 Nivedita.pdf (720.89 kB)
Download file

DEFENDING BERT AGAINST MISSPELLINGS

Download (720.89 kB)
thesis
posted on 06.04.2021, 13:53 by Nivedita NighojkarNivedita Nighojkar
Defending models against Natural Language Processing adversarial attacks is a challenge because of the discrete nature of the text dataset. However, given the variety of Natural Language Processing applications, it is important to make text processing models more robust and secure. This paper aims to develop techniques that will help text processing models such as BERT to combat adversarial samples that contain misspellings. These developed models are more robust than off the shelf spelling checkers.

History

Degree Type

Master of Science

Department

Computer and Information Technology

Campus location

West Lafayette

Advisor/Supervisor/Committee Chair

John Springer

Additional Committee Member 2

Jin Kocsis

Additional Committee Member 3

Eric Dietz

Usage metrics

Licence

Exports