SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2310.12508 · ICLR · 2024

SalUn: Empowering Machine Unlearning via Gradient-based Weight Saliency in Both Image Classification and Generation

Sijia Liu, Dennis Wei, Eric Wong, Yihua Zhang, Jiancheng Liu, Chongyu Fan

arXiv · PDF · Open in the Atlas

Code that ran

We lifted 11 functions out of this paper's own repositories and ran 8 of them in a sandbox. "Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results.

RepositoryRoleRan
optml-group/unlearn-saliency canonical 8 of 11
FunctionStatusWhere it lives
FGSM_perturb Ran optml-group/unlearn-saliency/Classification/unlearn/boundary_sh.py
code served (permissive licence) · get_code("074c8fc7870f5cc9")
discretize Ran optml-group/unlearn-saliency/Classification/unlearn/boundary_sh.py
code served (permissive licence) · get_code("20b8d1668895f8bb")
fisher Ran optml-group/unlearn-saliency/Classification/unlearn/fisher.py
code served (permissive licence) · get_code("3d980bf223d0e00b")
fisher_information_martix Ran optml-group/unlearn-saliency/Classification/unlearn/fisher.py
code served (permissive licence) · get_code("c25079d6cf22a5e2")
get_require_grad_params Ran optml-group/unlearn-saliency/Classification/unlearn/Wfisher.py
code served (permissive licence) · get_code("aebe5de64c1725b4")
l1_regularization Ran optml-group/unlearn-saliency/Classification/unlearn/FT.py
code served (permissive licence) · get_code("cae29c9fba744465")
sam_grad Ran optml-group/unlearn-saliency/Classification/unlearn/Wfisher.py
code served (permissive licence) · get_code("4145035921358f67")
woodfisher Ran optml-group/unlearn-saliency/Classification/unlearn/Wfisher.py
code served (permissive licence) · get_code("6fdded37bf329957")
GA Not yet run optml-group/unlearn-saliency/Classification/unlearn/GA_prune.py
code served (permissive licence) · get_code("1e7882ed2e1e9d82")
GA Not yet run optml-group/unlearn-saliency/Classification/unlearn/GA_prune_bi.py
code served (permissive licence) · get_code("552196cf57918ac6")
get_mean_var Not yet run optml-group/unlearn-saliency/Classification/unlearn/fisher.py
code served (permissive licence) · get_code("e953ece8832f74dd")

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

With evolving data regulations, machine unlearning (MU) has become an important tool for fostering trust and safety in today's AI models. However, existing MU methods focusing on data and/or weight perspectives often suffer limitations in unlearning accuracy, stability, and cross-domain applicability. To address these challenges, we introduce the concept of 'weight saliency' for MU, drawing parallels with input saliency in model explanation. This innovation directs MU's attention toward specific model weights rather than the entire model, improving effectiveness and efficiency. The resultant method that we call saliency unlearning (SalUn) narrows the performance gap with 'exact' unlearning (model retraining from scratch after removing the forgetting data points). To the best of our knowledge, SalUn is the first principled MU approach that can effectively erase the influence of forgetting data, classes, or concepts in both image classification and generation tasks. For example, SalUn yields a stability advantage in high-variance random data forgetting, e.g., with a 0.2% gap compared to exact unlearning on the CIFAR-10 dataset. Moreover, in preventing conditional diffusion models from generating harmful images, SalUn achieves nearly 100% unlearning accuracy, outperforming current state-ofthe-art baselines like Erased Stable Diffusion and Forget-Me-Not.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2310.12508")
get_code_for_paper("2310.12508")
have("2310.12508")

Connect an agent — have() is free.