Salvato in:
| Autori principali: | , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2024
|
| Soggetti: | |
| Accesso online: | https://arxiv.org/abs/2402.07689 |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866910401315209216 |
|---|---|
| author | Alekseevskaia, Irina Arkhipenko, Konstantin |
| author_facet | Alekseevskaia, Irina Arkhipenko, Konstantin |
| contents | The use of third-party datasets and pre-trained machine learning models poses a threat to NLP systems due to possibility of hidden backdoor attacks. Existing attacks involve poisoning the data samples such as insertion of tokens or sentence paraphrasing, which either alter the semantics of the original texts or can be detected. Our main difference from the previous work is that we use the reposition of a two words in a sentence as a trigger. By designing and applying specific part-of-speech (POS) based rules for selecting these tokens, we maintain high attack success rate on SST-2 and AG classification datasets while outperforming existing attacks in terms of perplexity and semantic similarity to the clean samples. In addition, we show the robustness of our attack to the ONION defense method. All the code and data for the paper can be obtained at https://github.com/alekseevskaia/OrderBkd. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2402_07689 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | OrderBkd: Textual backdoor attack through repositioning Alekseevskaia, Irina Arkhipenko, Konstantin Computation and Language Artificial Intelligence The use of third-party datasets and pre-trained machine learning models poses a threat to NLP systems due to possibility of hidden backdoor attacks. Existing attacks involve poisoning the data samples such as insertion of tokens or sentence paraphrasing, which either alter the semantics of the original texts or can be detected. Our main difference from the previous work is that we use the reposition of a two words in a sentence as a trigger. By designing and applying specific part-of-speech (POS) based rules for selecting these tokens, we maintain high attack success rate on SST-2 and AG classification datasets while outperforming existing attacks in terms of perplexity and semantic similarity to the clean samples. In addition, we show the robustness of our attack to the ONION defense method. All the code and data for the paper can be obtained at https://github.com/alekseevskaia/OrderBkd. |
| title | OrderBkd: Textual backdoor attack through repositioning |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2402.07689 |