Probing the Transition to Dataset-Level Privacy in ML Models Using an Output-Specific and Data-Resolved Privacy Profile.
Tyler LeBlondJoseph MunozFred LuMaya FuchsElliott Zaresky-WilliamsEdward RaffBrian TestaPublished in: AISec@CCS (2023)
Keyphrases
- data sets
- sensitive information
- data analysis
- privacy concerns
- private data
- training data
- privacy preserving
- input data
- data collection
- data privacy
- personal data
- privacy requirements
- image data
- sensitive data
- privacy protection
- private information
- database
- original data
- data quality
- differential privacy
- accurate models
- historical data
- prior knowledge
- experimental data
- synthetic data
- knowledge discovery
- preserving privacy
- differentially private
- data publishing
- probabilistic model
- probability distribution
- maximum likelihood
- data processing
- security issues
- personal information
- privacy preserving data mining
- model selection
- privacy issues
- data sources
- information loss
- information security
- data mining
- confidential data
- raw data