On the Robustness of Self-Attentive Models
By Yu-Lun Hsieh, Minhao Cheng, Da-Cheng Juan, Wei Wei, Wen-Lian Hsu, Cho-Jui Hsieh. 2019
110 citations recorded by the source corpus.
This work examines the robustness of self-attentive neural networks against adversarial input perturbations. Specifically, we investigate the attention and feature extraction mechanisms of state-of-the-art recurrent neural networks and self-attentive architectures for sentiment analysis, entailment and machine translation under adversarial attacks. We also propose a novel attack algorithm for generating more natural adversarial examples that could mislead neural models but not humans. Experimental results show that, compared to recurrent neural models, self-attentive models are more robust against adversarial perturbation. In addition, we provide theoretical explanations for their superior robustness to support our claims.
Subfields: Artificial Intelligence, Model Architectures, Safety, Robustness & Interpretability
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Artificial Intelligence
- Jianfeng Gao
- Weizhu Chen
- Xiaodong Liu
- Pengcheng He
- Samuel R. Bowman
- Mohit Bansal
- Yonatan Belinkov
- Percy Liang
In Model Architectures
- Weizhu Chen
- Ellie Pavlick
- Charles Lovering
- Marten van Dijk
- Kaleel Mahmood
- Hoifung Poon
- Jianfeng Gao
- Alan Nichol
In Safety, Robustness & Interpretability
- Charles Lovering
- Marten van Dijk
- Kaleel Mahmood
- Jianfeng Gao
- Xiaodong Liu
- Weizhu Chen
- Tal Linzen
- Ellie Pavlick
Citations in the corpus
- Adversarial Training for Large Neural Language Models (2020, background): “However, such models can still suffer catastrophic loss in adversarial scenarios (Nie et al., 2019; Hsieh et al., 2019; Madry et al., 2017; Jin et al., 2019), with attacks as simple as replacing a few words in input sentences while preserving the semantics.”
- On the Robustness of Vision Transformers to Adversarial Examples (2021, background): “The transformer has been well studied from an adversarial perspective for NLP applications e.g., [18, 31, 17, 15].”
- Theoretical Limitations of Self-Attention in Neural Sequence Models (2019, background): “A range of studies has experimentally examined the ability of recurrent networks to model counter languages such as a n b n (Kalinke and Lehmann, 1998; Gers and Schmidhuber, 2001; Cartling, 2008; Weiss et al., 2018; Suzgun et al., 2019).”
- Predicting Inductive Biases of Pre-Trained Models (2021, background): “Iyyer et al., 2018; Hsieh et al., 2019; Jia et al., 2019; Alzantot et al., 2018; Hsieh et al., 2019; Ilyas et al., 2019; Madry et al., 2017; Athalye et al., 2018) is also relevant, as it relates to the influence of dataset artifacts on models’ decisions.”
- Dialogue Transformers (2019, background): “In addition, transformers have recently shown to be more robust to unexpected inputs (such as adversarial examples) [13].”