Synthetic Query Generation for Privacy-Preserving Deep Retrieval Systems using Differentially Private Language Models.
Aldo G. CarranzaRezsa FarahaniNatalia PonomarevaAlexey KurakinMatthew JagielskiMilad NasrPublished in: NAACL-HLT (2024)
Keyphrases
- query generation
- language model
- retrieval systems
- differentially private
- privacy preserving
- differential privacy
- test collection
- information retrieval
- language modeling
- retrieval model
- information retrieval systems
- privacy preservation
- query expansion
- text retrieval
- privacy guarantees
- average precision
- privacy preserving data mining
- probabilistic model
- multimedia
- document retrieval
- retrieval effectiveness
- relevant documents
- search engine
- sensitive information
- data privacy
- query logs
- privacy concerns
- evaluation measures
- pseudo relevance feedback
- ranked list
- learning to rank
- sensitive data
- query terms
- database
- privacy protection
- decision trees
- knowledge discovery