Metraipolis › archive › dc_db6cc495bb96

Data Unlearning via Inverse Distillation — arXiv:2609.36099

Source: arXiv (en)
Original: https://arxiv.org/abs/2609.36099
License (as archived): CC0-1.0
Archived: 2026-10-07T17:28:42.207Z
SHA-256 of the archived text: ae328843edeedcc8f98f810c68a43917ecdbd37563bc9a0c364dfd64842113b9
Stable URL of this page: https://metraipolis.com/documento/dc_db6cc495bb96

How to cite

Data Unlearning via Inverse Distillation — arXiv:2609.36099 — arXiv (en), https://arxiv.org/abs/2609.36099 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_db6cc495bb96 (Metraipolis archive dc_db6cc495bb96, sha256 ae328843edeedcc8f98f810c68a43917ecdbd37563bc9a0c364dfd64842113b9)

BibTeX

@misc{dcdb6cc495bb96,
  title = {Data Unlearning via Inverse Distillation — arXiv:2609.36099},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.36099},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_db6cc495bb96; sha256 ae328843edeedcc8f98f810c68a43917ecdbd37563bc9a0c364dfd64842113b9},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.36099. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Multi-step matching models, including flow and diffusion models, produce high-quality outputs but incur substantial inference costs and may reproduce unwanted components of their training datasets. We introduce Inverse Distillation Unlearning (IDU), a unified framework that simultaneously distills a teacher multi-step matching model into an efficient one-step student generator and suppresses outputs corresponding to a designated training subset. We first formulate distillation as a min-max objective over a data distribution and then represent this distribution as a mixture of the forget-set and the generated distributions. This allows us to compare this mixture with the teacher's training distribution and recover only the retained data at the optimum. Our method requires only a pretrained full-data teacher and data from the forget set, without access to retained training examples, extra feature extractors or classifiers. Extensive experiments on MNIST and CIFAR-10 datasets under flow-matching and score-based diffusion settings demonstrate that IDU substantially reduces the generation frequency of forgotten classes while preserving generation quality on the retained classes. To the best of our knowledge, IDU is the first unified framework for simultaneous unlearning and distillation in unconditional flow-matching and score-based models.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-30 in cs.LG).
Categorie arXiv: Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2609.36099 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-07T16:59:48.099Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-09-30?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).