PPL
Alina Oprea
Papers (6)
PoisonForge: Task-Level Targeted Poisoning Benchmark for Instruction-Tuned LLMsWho Owns This Agent? Tracing AI Agents Back to Their OwnersAPWA: A Distributed Architecture for Parallelizable Agentic WorkflowsAttacks and Mitigations for Distributed Governance of Agentic AI under Byzantine AdversariesReconstruction of Personally Identifiable Information from Supervised Finetuned ModelsMAGIQ: A Post-Quantum Multi-Agentic AI Governance System with Provable Security