The Frontier of Data Erasure: A Survey on Machine Unlearning for Large Language Models

Youyang Qu, Ming Ding, Nan Sun, Kanchana Thilakarathna, Tianqing Zhu, Dusit Niyato

Abstract: This article reviews the latest in machine unlearning for large language models (LLMs), introducing methods for the targeted forgetting of information to address privacy, ethical, and legal challenges without necessitating full-model retraining. Machine unlearning is a cutting-edge solution, mitigating risk concerns of LLMs and artificial intelligence.