Agentic Unlearning: When LLM Agent Meets Machine Unlearning
This addresses the problem of sensitive data removal in AI agents for domains like healthcare, though it is incremental as it builds on existing unlearning methods by extending them to memory pathways.
The paper tackles the problem of removing specified information from both model parameters and persistent memory in agents with closed-loop interaction, introducing agentic unlearning and the Synchronized Backflow Unlearning (SBU) framework, which reduces traces of targeted private information across both pathways with limited degradation on retained data in experiments on medical QA benchmarks.
In this paper, we introduce \textbf{agentic unlearning} which removes specified information from both model parameters and persistent memory in agents with closed-loop interaction. Existing unlearning methods target parameters alone, leaving two critical gaps: (i) parameter-memory backflow, where retrieval reactivates parametric remnants or memory artifacts reintroduce sensitive content, and (ii) the absence of a unified strategy that covers both parameter and memory pathways. We present Synchronized Backflow Unlearning (SBU), a framework that unlearns jointly across parameter and memory pathways. The memory pathway performs dependency closure-based unlearning that prunes isolated entities while logically invalidating shared artifacts. The parameter pathway employs stochastic reference alignment to guide model outputs toward a high-entropy prior. These pathways are integrated via a synchronized dual-update protocol, forming a closed-loop mechanism where memory unlearning and parametric suppression reinforce each other to prevent cross-pathway recontamination. Experiments on medical QA benchmarks show that SBU reduces traces of targeted private information across both pathways with limited degradation on retained data.