[연구] 이지형 교수 연구실, EMNLP 2022 논문 게재 승인
- 소프트웨어융합대학
- 조회수1151
- 2022-10-25
이지형 교수 연구실, EMNLP 2022 논문 게재 승인
정보 및 지능 시스템 연구실(지도교수: 이지형)의 최윤석(소프트웨어학과 박사과정)과 김효준(인공지능학과 석사과정)의 "TABS: Efficient Textual Adversarial Attack for Pre-trained NL Code Model Using Semantic Beam Search" 논문이 세계 최고 권위의 자연어처리분야 학회 'EMNLP(Empirical Methods in Natural Language Processing) 2022'에 게재 승인되었습니다. 2022년 12월 UAE 아부다비에서 발표될 예정입니다.
YunSeok Choi, Hyojun Kim, and Jee-Hyong Lee. "TABS: Efficient Textual Adversarial Attack for Pre-trained NL Code Model Using Semantic Beam Search" In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing: EMNLP 2022, Abu Dhabi, UAE, 2022
본 연구에서는 beam search 기반의 black-box adversarial attack method인 TABS를 제안합니다. 기존 연구는 간단한 greedy search 기반의 비효율적인 방법으로 adversarial example을 생성하지만, TABS는 beam search와 contextual semantic filtering를 통해 더 나은 adversarial example을 생성함과 동시에 search space를 효율적으로 줄이는 것을 확인하였습니다. 제안된 방법론은 NL code search classification과 retrieval task에서 attack success, the number of queires, semantic similarity 모두 향상된 성능을 달성하였습니다.
Abstract: As pre-trained models have shown successful performance in program language processing as well as natural language processing, adversarial attacks on these models also attract attention. However, previous works on blackbox adversarial attacks generated adversarial examples in a very inefficient way with simple greedy search. They also failed to find out better adversarial examples because it was hard to reduce the search space without performance loss. In this paper, we propose TABS, an efficient beam search black-box adversarial attack method. We adopt beam search to find out better adversarial examples, and contextual semantic filtering to effectively reduce the search space. Contextual semantic filtering reduces the number of candidate adversarial words considering the surrounding context and the semantic similarity. Our proposed method shows good performance in terms of attack success rate, the number of queries, and semantic similarity in attacking models for two tasks: NL code search classification and retrieval tasks.