When Does Machine Learning FAIL? Generalized Transferability for Evasion and Poisoning Attacks.
Octavian SuciuRadu MargineanYigitcan KayaHal Daumé IIITudor DumitrasPublished in: USENIX Security Symposium (2018)
Keyphrases
- machine learning
- countermeasures
- machine learning algorithms
- machine learning methods
- explanation based learning
- information security
- computer security
- support vector machine
- learning algorithm
- digital images
- natural language processing
- text mining
- knowledge acquisition
- malicious attacks
- computer vision
- learning systems
- artificial intelligence
- machine learning approaches
- inductive learning
- traffic analysis
- learning problems
- learning tasks
- security risks
- denial of service attacks
- computational intelligence
- supervised learning
- information extraction
- pattern recognition
- chosen plaintext
- spam filters
- security protocols
- spam filtering
- network security
- inductive logic programming
- data mining
- information systems
- semi supervised learning
- active learning