Enhance Robustness of Language Models Against Variation Attack through Graph Integration
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2024
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866916212086145024 |
|---|---|
| author | Xiong, Zi Qing, Lizhi Kang, Yangyang Liu, Jiawei Li, Hongsong Sun, Changlong Liu, Xiaozhong Lu, Wei |
| author_facet | Xiong, Zi Qing, Lizhi Kang, Yangyang Liu, Jiawei Li, Hongsong Sun, Changlong Liu, Xiaozhong Lu, Wei |
| contents | The widespread use of pre-trained language models (PLMs) in natural language processing (NLP) has greatly improved performance outcomes. However, these models' vulnerability to adversarial attacks (e.g., camouflaged hints from drug dealers), particularly in the Chinese language with its rich character diversity/variation and complex structures, hatches vital apprehension. In this study, we propose a novel method, CHinese vAriatioN Graph Enhancement (CHANGE), to increase the robustness of PLMs against character variation attacks in Chinese content. CHANGE presents a novel approach for incorporating a Chinese character variation graph into the PLMs. Through designing different supplementary tasks utilizing the graph structure, CHANGE essentially enhances PLMs' interpretation of adversarially manipulated text. Experiments conducted in a multitude of NLP tasks show that CHANGE outperforms current language models in combating against adversarial attacks and serves as a valuable contribution to robust language model research. These findings contribute to the groundwork on robust language models and highlight the substantial potential of graph-guided pre-training strategies for real-world applications. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2404_12014 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Enhance Robustness of Language Models Against Variation Attack through Graph Integration Xiong, Zi Qing, Lizhi Kang, Yangyang Liu, Jiawei Li, Hongsong Sun, Changlong Liu, Xiaozhong Lu, Wei Computation and Language Cryptography and Security The widespread use of pre-trained language models (PLMs) in natural language processing (NLP) has greatly improved performance outcomes. However, these models' vulnerability to adversarial attacks (e.g., camouflaged hints from drug dealers), particularly in the Chinese language with its rich character diversity/variation and complex structures, hatches vital apprehension. In this study, we propose a novel method, CHinese vAriatioN Graph Enhancement (CHANGE), to increase the robustness of PLMs against character variation attacks in Chinese content. CHANGE presents a novel approach for incorporating a Chinese character variation graph into the PLMs. Through designing different supplementary tasks utilizing the graph structure, CHANGE essentially enhances PLMs' interpretation of adversarially manipulated text. Experiments conducted in a multitude of NLP tasks show that CHANGE outperforms current language models in combating against adversarial attacks and serves as a valuable contribution to robust language model research. These findings contribute to the groundwork on robust language models and highlight the substantial potential of graph-guided pre-training strategies for real-world applications. |
| title | Enhance Robustness of Language Models Against Variation Attack through Graph Integration |
| topic | Computation and Language Cryptography and Security |
| url | https://arxiv.org/abs/2404.12014 |