OFMU：面向机器遗忘的优化驱动框架
OFMU: Optimization-Driven Framework for Machine Unlearning
AI总结：
提出OFMU，一种基于惩罚的双层优化框架，通过层次结构优先遗忘并保持保留性能，在视觉和语言基准上优于现有方法。
摘要：
部署在敏感应用中的大型语言模型越来越需要具备遗忘特定知识的能力，例如用户请求、受版权保护的材料或过时信息，而无需从头重新训练，以确保法规合规性、用户隐私和安全。这一任务称为机器遗忘，旨在移除目标数据的影响（遗忘），同时保持对剩余数据的性能（保留）。一种常见方法是将此表述为多目标问题，并通过标量化将其简化为单目标问题，其中遗忘损失和保留损失通过加权和组合。然而，由于梯度方向冲突，这通常会导致不稳定的训练动态和模型效用下降。为了解决这些挑战，我们提出了OFMU，一种基于惩罚的双层优化框架，通过层次结构明确优先遗忘同时保持保留。我们的方法通过内部最大化步骤强制执行遗忘，该步骤结合了相似性感知惩罚以解耦遗忘和保留目标的梯度，并通过外部最小化步骤恢复效用。为了确保可扩展性，我们开发了一种双循环算法，在凸和非凸情况下均具有可证明的收敛保证。我们进一步提供了收敛率的严格理论分析，并表明我们的方法在遗忘效果和模型效用之间实现了比先前方法更好的权衡。在视觉和语言基准上的大量实验表明，OFMU在遗忘效果和保留效用方面均持续优于现有的遗忘方法。
英文摘要：
Large language models deployed in sensitive applications increasingly require the ability to unlearn specific knowledge, such as user requests, copyrighted materials, or outdated information, without retraining from scratch to ensure regulatory compliance, user privacy, and safety. This task, known as machine unlearning, aims to remove the influence of targeted data (forgetting) while maintaining performance on the remaining data (retention). A common approach is to formulate this as a multi-objective problem and reduce it to a single-objective problem via scalarization, where forgetting and retention losses are combined using a weighted sum. However, this often results in unstable training dynamics and degraded model utility due to conflicting gradient directions. To address these challenges, we propose OFMU, a penalty-based bi-level optimization framework that explicitly prioritizes forgetting while preserving retention through a hierarchical structure. Our method enforces forgetting via an inner maximization step that incorporates a similarity-aware penalty to decorrelate the gradients of the forget and retention objectives, and restores utility through an outer minimization step. To ensure scalability, we develop a two-loop algorithm with provable convergence guarantees under both convex and non-convex regimes. We further provide a rigorous theoretical analysis of convergence rates and show that our approach achieves better trade-offs between forgetting efficacy and model utility compared to prior methods. Extensive experiments across vision and language benchmarks demonstrate that OFMU consistently outperforms existing unlearning methods in both forgetting efficacy and retained utility.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
