TRAPDOOR: Repurposing neural network backdoors to detect dataset bias in machine learning-based genomic analysis.
Esha SarkarConstantine DoumanidisMichail ManiatakosPublished in: VLSI-SoC (2023)
Keyphrases
- neural network
- machine learning
- data sets
- artificial neural networks
- computer vision
- pattern recognition
- statistical analysis
- genetic algorithm
- np complete
- satisfiability problem
- recurrent neural networks
- benchmark datasets
- detection algorithm
- intrusion detection
- support vector machine
- active learning
- feature selection
- artificial intelligence