ParaFuzz: An Interpretability-Driven Technique for Detecting Poisoned Samples in NLP.
Lu YanZhuo ZhangGuanhong TaoKaiyuan ZhangXuan ChenGuangyu ShenXiangyu ZhangPublished in: NeurIPS (2023)
Keyphrases
- natural language processing
- data sets
- question answering
- text analysis
- data driven
- prediction accuracy
- language processing
- automatic detection
- machine translation
- text processing
- text mining
- natural language
- database
- training set
- semantic analysis
- training dataset
- part of speech
- case study
- artificial intelligence
- computational linguistics
- sampling methods
- field of natural language processing