Token-Modification Adversarial Attacks for Natural Language Processing: A Survey
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2021
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866911749529141248 |
|---|---|
| author | Roth, Tom Gao, Yansong Abuadbba, Alsharif Nepal, Surya Liu, Wei |
| author_facet | Roth, Tom Gao, Yansong Abuadbba, Alsharif Nepal, Surya Liu, Wei |
| contents | Many adversarial attacks target natural language processing systems, most of which succeed through modifying the individual tokens of a document. Despite the apparent uniqueness of each of these attacks, fundamentally they are simply a distinct configuration of four components: a goal function, allowable transformations, a search method, and constraints. In this survey, we systematically present the different components used throughout the literature, using an attack-independent framework which allows for easy comparison and categorisation of components. Our work aims to serve as a comprehensive guide for newcomers to the field and to spark targeted research into refining the individual attack components. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2103_00676 |
| institution | arXiv |
| publishDate | 2021 |
| record_format | arxiv |
| spellingShingle | Token-Modification Adversarial Attacks for Natural Language Processing: A Survey Roth, Tom Gao, Yansong Abuadbba, Alsharif Nepal, Surya Liu, Wei Computation and Language Cryptography and Security Machine Learning Many adversarial attacks target natural language processing systems, most of which succeed through modifying the individual tokens of a document. Despite the apparent uniqueness of each of these attacks, fundamentally they are simply a distinct configuration of four components: a goal function, allowable transformations, a search method, and constraints. In this survey, we systematically present the different components used throughout the literature, using an attack-independent framework which allows for easy comparison and categorisation of components. Our work aims to serve as a comprehensive guide for newcomers to the field and to spark targeted research into refining the individual attack components. |
| title | Token-Modification Adversarial Attacks for Natural Language Processing: A Survey |
| topic | Computation and Language Cryptography and Security Machine Learning |
| url | https://arxiv.org/abs/2103.00676 |