Second-Order Information Matters: Revisiting Machine Unlearning for Large Language Models
This work addresses privacy and copyright concerns for LLM practitioners by providing a more robust and efficient unlearning method, though it is incremental as it builds on existing gradient-based approaches.
The paper tackles the problem of machine unlearning for Large Language Models to address privacy and copyright issues by proposing algorithms based on second-order information (Hessian), which outperform first-order methods in utility preservation and privacy guarantees across multiple NLP datasets.
With the rapid development of Large Language Models (LLMs), we have witnessed intense competition among the major LLM products like ChatGPT, LLaMa, and Gemini. However, various issues (e.g. privacy leakage and copyright violation) of the training corpus still remain underexplored. For example, the Times sued OpenAI and Microsoft for infringing on its copyrights by using millions of its articles for training. From the perspective of LLM practitioners, handling such unintended privacy violations can be challenging. Previous work addressed the ``unlearning" problem of LLMs using gradient information, while they mostly introduced significant overheads like data preprocessing or lacked robustness. In this paper, contrasting with the methods based on first-order information, we revisit the unlearning problem via the perspective of second-order information (Hessian). Our unlearning algorithms, which are inspired by classic Newton update, are not only data-agnostic/model-agnostic but also proven to be robust in terms of utility preservation or privacy guarantee. Through a comprehensive evaluation with four NLP datasets as well as a case study on real-world datasets, our methods consistently show superiority over the first-order methods.