ZK-APEX is a zero-shot verifiable approximate unlearning framework that applies sparse masking and blockwise empirical Fisher compensation on personalized models and uses Halo2 proofs to confirm correct execution without revealing client data.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CR 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
ZK-APEX: Zero-Knowledge Approximate Personalized Unlearning with Executable Proofs
ZK-APEX is a zero-shot verifiable approximate unlearning framework that applies sparse masking and blockwise empirical Fisher compensation on personalized models and uses Halo2 proofs to confirm correct execution without revealing client data.