Hsiang Hsu, Chun-Fu Chen, Ivan Brugere, Freddy Lecue, Pradeep Niroula, Zichang He
We lifted 11 functions out of this paper's own repositories and ran 1 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.
| Repository | Role | Ran |
|---|---|---|
| Graph-COM/Langevin_unlearning | canonical | 1 of 5 |
| joeljang/knowledge-unlearning | canonical | 0 of 2 |
| AdityaGolatkar/SelectiveForgetting | — | 0 of 4 |
| Function | Status | Where it lives |
|---|---|---|
| unadjusted_langevin_algorithm | Ran | Graph-COM/Langevin_unlearning/langevin.py pointer only (licence: NONE) · get_code("d14dd84200ac6432") |
| _add_logvar | Not yet run | AdityaGolatkar/SelectiveForgetting/variational.py pointer only (licence: NOASSERTION) · get_code("23ab3aca3edc9e96") |
| _reset_logvar | Not yet run | AdityaGolatkar/SelectiveForgetting/variational.py pointer only (licence: NOASSERTION) · get_code("e37896de279105a1") |
| load_features | Not yet run | Graph-COM/Langevin_unlearning/utils.py pointer only (licence: NONE) · get_code("b149132fab16d2fd") |
| make_variational | Not yet run | AdityaGolatkar/SelectiveForgetting/variational.py pointer only (licence: NOASSERTION) · get_code("cde9bb694fbfc040") |
| normalize_reply | Not yet run | joeljang/knowledge-unlearning/utils.py pointer only (licence: NONE) · get_code("6e5d38d011c3c8dc") |
| onehot | Not yet run | Graph-COM/Langevin_unlearning/utils.py pointer only (licence: NONE) · get_code("424a51b62f7dd524") |
| transform_array | Not yet run | Graph-COM/Langevin_unlearning/utils.py pointer only (licence: NONE) · get_code("46155c29dca22749") |
| unadjusted_langevin_algorithm_multiclass | Not yet run | Graph-COM/Langevin_unlearning/langevin.py pointer only (licence: NONE) · get_code("0ba8166fbd1eea1c") |
| uppercase | Not yet run | joeljang/knowledge-unlearning/utils.py pointer only (licence: NONE) · get_code("b3ed4c3bccc10c65") |
| variational_forward | Not yet run | AdityaGolatkar/SelectiveForgetting/variational.py pointer only (licence: NOASSERTION) · get_code("e16667b5fa8b0848") |
Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.
Machine unlearning offers a practical alternative to avoid full model re-training by approximately removing the influence of specific user data. While existing methods certify unlearning via statistical indistinguishability from re-trained models, these guarantees do not naturally extend to model outputs when inputs are adversarially perturbed. In particular, slight perturbations of forget samples may still be correctly recognized by the unlearned model-even when a re-trained model fails to do so-revealing a novel privacy risk: information about the forget samples may persist in their local neighborhood. In this work, we formalize this vulnerability as residual knowledge and show that it is inevitable in high-dimensional settings. To mitigate this risk, we propose a fine-tuning strategy, named RURK, that penalizes the model's ability to re-recognize perturbed forget samples. Experiments on vision benchmarks with deep neural networks demonstrate that residual knowledge is prevalent across existing unlearning methods and that our approach effectively prevents residual knowledge.
The same record, over MCP at https://syntology.ai/mcp:
get_harvested_code_for_paper("2601.22359")
get_code_for_paper("2601.22359")
have("2601.22359")
Connect an agent — have() is free.