EMNLP Findings2022
NaturalAdversaries: Can Naturalistic Adversaries Be as Effective as Artificial Adversaries?
Abstract
While a substantial body of prior work has explored adversarial example generation for natural language understanding tasks, these examples are often unrealistic and diverge from the real-world data distributions. In this work, we introduce a two-stage adversarial example generation framework (NaturalAdversaries), for designing adversaries that are effective at fooling a given classifier and demonstrate naturallooking failure cases that could plausibly occur during in-the-wild deployment of the models.
At the first stage a token attribution method is used to summarize a given classifier's behaviour as a function of the key tokens in the input. In the second stage a generative model is conditioned on the key tokens from the first stage. NaturalAdversaries is adaptable to both black-box and white-box adversarial attacks based on the level of access to the model parameters. Our results indicate these adversaries generalize across domains, and offer insights for future research on improving robustness of neural text classification models.
Ask Lune
Know this paper in minutes, not an afternoon.
Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.
Lineage
Builds on15
- The Curious Case of Neural Text DegenerationAri Holtzman, Jan Buys, Li Du, Maxwell Forbes et al.ICLR 2020 · 4,112 citations
- Deberta: decoding-Enhanced Bert with Disentangled AttentionPengcheng He, Xiaodong Liu, Jianfeng Gao, Weizhu ChenICLR 2021 · 3,729 citations
- Is BERT Really Robust? A Strong Baseline for Natural Language Attack on Text Classification and EntailmentDi Jin, Zhijing Jin, Joey Tianyi Zhou, Peter SzolovitsAAAI 2020 · 1,333 citations
- Adversarial NLI: A New Benchmark for Natural Language UnderstandingYixin Nie, Adina Williams, Emily Dinan, Mohit Bansal et al.ACL 2020 · 602 citations
- Adversarial Filters of Dataset BiasesRonan Le Bras, Swabha Swayamdipta, Chandra Bhagavatula, Rowan Zellers et al.ICML 2020 · 242 citations
Related papers
- Natural Adversarial ExamplesDan Hendrycks, Kevin Zhao, Steven Basart, Jacob Steinhardt et al.CVPR 2021
- Fooling the Textual Fooler via Randomizing Latent RepresentationsDuy C. Hoang, Nguyen Hung-Quang, Saurav Manchanda, Minlong Peng et al.ACL Findings 2024 · 1 citation
- Fooling Explanations in Text ClassifiersAdam Ivankay, Ivan Girardi, Chiara Marchiori, Pascal FrossardICLR 2022 · 23 citations
- Randomized Smoothing with Masked Inference for Adversarially Robust Text ClassificationsHan Cheol Moon, Shafiq R. Joty, Ruochen Zhao, Megh Thakkar et al.ACL 2023 · 6 citations
- Identifying Human Strategies for Generating Word-Level Adversarial ExamplesMaximilian Mozes, Bennett Kleinberg, Lewis D. GriffinEMNLP Findings 2022 · 1 citation
