kartikeya sharma
Differential privacy · Applied cryptography · Verifiable computation

Differential privacy, applied cryptography, and what breaks when the accounting is not checked.

Protocol Review of SecureNN

A review of eprint 2018/442 checked against both published revisions and the reference implementation. Five questions, one real gap, and a routing typo in the 4-party protocol that hands one server the plaintext.

Secure Multi-Party ComputationProtocol ReviewPrivacy Preserving Machine Learning
Read full article →

How SMOTE Quietly Cancels Your Differential Privacy

Oversampling before private training puts each record and its synthetic copies in one correlated group, so a record-level guarantee silently becomes a group-level one and the epsilon you report stops meaning what it claims.

Differential PrivacyAdaptive DataMachine Learning
Read full article →

What It Takes To Build A Differentially Private Mechanism

Sensitivity, composition, and clipping, worked end to end. What has to hold before a privacy budget is a guarantee rather than a number in a config file.

Differential PrivacyAdaptive IntelligenceEfficiency and Adaptive Compute
Read full article →

3 Models for Formal Verification of Micropayment Agent Guardrails

Three verification models for autonomous agent payment guardrails under X402 and AP2, covering transaction neutrality, active verification oracles, and adversarial resilience against planted backdoors.

AI SafetyFormal VerificationCryptography
Read full article →

Formal Verification of LLM Safety with Zero Knowledge Proofs

What a zero-knowledge proof can and cannot establish about a refusal. Proofs give integrity, and a safety failure is usually not an integrity problem.

AI SafetyZero Knowledge ProofsLLM Security
Read full article →