Universal Adversarial Triggers for Attacking and Analyzing NLP
Explore this paper's citation graph
- Type
- preprint
- Published
- 2019-08-20
- Cited by
- 1,130
- References
- 38
- Access
- Open access
- OpenAlex
- https://openalex.org/W2970290563
- Semantic Scholar
- https://api.semanticscholar.org/CorpusID:201698258
Keywords
Computer science, Adversarial system, Heuristics, Artificial intelligence, Language model
References
- Intriguing properties of neural networks
- Neural Machine Translation of Rare Words with Subword Units
- A large annotated corpus for learning natural language inference
- 2005 Special Issue: Framewise phoneme classification with bidirectional LSTM and other neural network architectures
- GloVe: Global Vectors for Word Representation
- Recursive Deep Models for Semantic Compositionality Over a Sentiment Treebank
- Crafting adversarial input sequences for recurrent neural networks
- A Decomposable Attention Model for Natural Language Inference
- Universal Adversarial Perturbations
- Understanding Neural Networks through Representation Erasure
- Enhanced LSTM for Natural Language Inference
- HotFlip: White-Box Adversarial Examples for Text Classification
- Deep Contextualized Word Representations
- Annotation Artifacts in Natural Language Inference Data
- Adversarial Example Generation with Syntactically Controlled Paraphrase Networks
- Hypothesis Only Baselines in Natural Language Inference
- Semantically Equivalent Adversarial Rules for Debugging NLP models
- Adversarial Patch
- On Adversarial Examples for Character-Level Neural Machine Translation
- Pathologies of Neural Models Make Interpretations Difficult
Cited by
- Hierarchical Reinforcement Learning for Open-Domain Dialog
- Classification as Decoder: Trading Flexibility for Control in Medical Dialogue
- What does BERT Learn from Multiple-Choice Reading Comprehension Datasets?
- Reducing Sentiment Bias in Language Models via Counterfactual Evaluation
- Negated LAMA: Birds cannot fly
- Adversarial Language Games for Advanced Natural Language Intelligence
- Word-level Textual Adversarial Attacking as Combinatorial Optimization
- TX-Ray: Quantifying and Explaining Model-Knowledge Transfer in (Un-)Supervised NLP
- How Can We Know What Language Models Know?
- WaLDORf: Wasteless Language-model Distillation On Reading-comprehension
- Learning the Difference that Makes a Difference with Counterfactually-Augmented Data
- AdvCodec: Towards A Unified Framework for Adversarial Text Generation
- Plug and Play Language Models: A Simple Approach to Controlled Text Generation
- Getting Closer to AI Complete Question Answering: A Set of Prerequisite Real Tasks
- Humpty Dumpty: Controlling Word Meanings via Corpus Poisoning
- Theory In, Theory Out: The Uses of Social Theory in Machine Learning for Social Science
- A Primer in BERTology: What We Know About How BERT Works
- From static to dynamic word representations: a survey
- Toward Interpretability of Dual-Encoder Models for Dialogue Response Suggestions
- Pre-trained models for natural language processing: A survey
Related papers
- A Context Aware Approach for Generating Natural Language Attacks
- Generating Semantically Valid Adversarial Questions for TableQA
- A View on Vulnerabilites: The Security Challenges of XAI (Academic Track)
- Event Detection without Triggers
- Are you tough enough? Framework for Robustness Validation of Machine Comprehension Systems
- Event Detection with Trigger-Aware Lattice Neural Network
- Zero-shot Label-Aware Event Trigger and Argument Classification
- Misleading Sentiment Analysis: Generating Adversarial Texts by the Ensemble Word Addition Algorithm
- Leveraging Dependency Regularization for Event Extraction