“SOPHON: Non-Fine-Tunable Learning to Restrain Task Transferability For Pre-Trained Models”, Jiangyi Deng, Shengyuan Pang, Yanjiao Chen, Liangming Xia, Yijie Bai, Haiqin Weng, Wenyuan Xu2024-04-19 (meta-learning, AI safety)⁠:

[previously] Instead of building deep learning models from scratch, developers are more and more relying on adapting pre-trained models to their customized tasks. However, powerful pre-trained models may be misused for unethical or illegal tasks, eg. privacy inference and unsafe content generation. In this paper, we introduce a pioneering learning paradigm, non-fine-tunable learning, which prevents the pre-trained model from being fine-tuned to indecent tasks while preserving its performance on the original task.

To fulfill this goal, we propose SOPHON, a protection framework that reinforces a given pre-trained model to be resistant to being fine-tuned in pre-defined restricted domains. Nonetheless, this is challenging due to a diversity of complicated fine-tuning strategies that may be adopted by adversaries. Inspired by model-agnostic meta-learning, we overcome this difficulty by designing sophisticated fine-tuning simulation and fine-tuning evaluation algorithms. In addition, we carefully design the optimization process to entrap the pre-trained model within a hard-to-escape local optimum regarding restricted domains.

We have conducted extensive experiments on two deep learning modes (classification and generation), 7 restricted domains, and 6 model architectures to verify the effectiveness of SOPHON.

Experiment results verify that fine-tuning SOPHON-protected models incurs an overhead comparable to or even greater than training from scratch. Furthermore, we confirm the robustness of SOPHON to 3 fine-tuning methods, 5 optimizers, various learning rates, and batch sizes.

SOPHON may help boost further investigations into safe and responsible AI.