SecureLLM: Using Compositionality to Build Provably Secure Language Models for Private, Sensitive, and Secret Data.
Abdulrahman AlabdulakreemChristian M. ArnoldYerim LeePieter M. FeenstraBoris KatzAndrei BarbuPublished in: CoRR (2024)
Keyphrases
- language model
- provably secure
- secret data
- secret key
- key management
- data hiding
- diffie hellman
- public key
- hash functions
- signature scheme
- standard model
- data embedding
- security model
- probabilistic model
- information retrieval
- mixture model
- smart card
- encryption scheme
- grayscale images
- pixel values
- error correction
- security analysis
- image compression
- low cost