Metraipolis › archive › dc_0eadee7d7b60

Improving Generative Model Self-Training with Geometrically Modified Outputs — arXiv:2609.35512

Source: arXiv (en)
Original: https://arxiv.org/abs/2609.35512
License (as archived): CC0-1.0
Archived: 2026-10-07T21:03:42.344Z
SHA-256 of the archived text: 25d5353ddc69b8038a9d05981be41703f9739504bb387e7780b47440a630f48b
Stable URL of this page: https://metraipolis.com/documento/dc_0eadee7d7b60

How to cite

Improving Generative Model Self-Training with Geometrically Modified Outputs — arXiv:2609.35512 — arXiv (en), https://arxiv.org/abs/2609.35512 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_0eadee7d7b60 (Metraipolis archive dc_0eadee7d7b60, sha256 25d5353ddc69b8038a9d05981be41703f9739504bb387e7780b47440a630f48b)

BibTeX

@misc{dc0eadee7d7b60,
  title = {Improving Generative Model Self-Training with Geometrically Modified Outputs — arXiv:2609.35512},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.35512},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_0eadee7d7b60; sha256 25d5353ddc69b8038a9d05981be41703f9739504bb387e7780b47440a630f48b},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.35512. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Self-training generative models - the continued improvement of a model using its own outputs - is becoming increasingly important as high-quality training data becomes scarce. However, naively finetuning on model-generated samples leads to degradation through model collapse and the model autophagy disorder. Negative-guidance self-training methods turn this degradation into a useful signal, using a model finetuned on its own outputs to guide the original model toward improved generation. Existing methods, however, take the negative signal in standard model outputs as given. We instead ask whether this signal can be explicitly strengthened. We introduce Geometrically Modified Outputs (GMOs), which reweight the singular values of the generator's input-output Jacobian to increase the influence of its leading singular directions. This geometric modification amplifies the mode-seeking behavior and distortions of standard outputs, providing a stronger and more targeted negative signal for self-training. Across a range of one-step generative models, GMOs consistently improve the performance of negative-guidance methods, including Neon and SIMS, compared with using standard model outputs.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-29 in cs.AI).
Categorie arXiv: Machine Learning (cs.LG) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2609.35512 — 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 2026-10-07T20:29:03.379Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-29?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).