Metraipolis › archive › dc_4aa97f590692

Broken scale symmetries in undercomplete linear autoencoders — arXiv:2610.03640

Source
arXiv
Original
https://arxiv.org/abs/2610.03640
License (as archived)
CC0-1.0
Archived
2026-10-05T23:10:26.326Z
SHA-256 of the archived text
93a5dbaf9ca9b57f9be4ce34f4929fb961f2cb5c9085ed15b05ac094074a8bdf
Stable URL of this page
https://metraipolis.com/documento/dc_4aa97f590692

How to cite

Broken scale symmetries in undercomplete linear autoencoders — arXiv:2610.03640 — arXiv, https://arxiv.org/abs/2610.03640 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_4aa97f590692 (Metraipolis archive dc_4aa97f590692, sha256 93a5dbaf9ca9b57f9be4ce34f4929fb961f2cb5c9085ed15b05ac094074a8bdf)
BibTeX
@misc{dc4aa97f590692,
  title = {Broken scale symmetries in undercomplete linear autoencoders — arXiv:2610.03640},
  howpublished = {arXiv},
  url = {https://arxiv.org/abs/2610.03640},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_4aa97f590692; sha256 93a5dbaf9ca9b57f9be4ce34f4929fb961f2cb5c9085ed15b05ac094074a8bdf},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv, https://arxiv.org/abs/2610.03640. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Neural network loss landscapes have many symmetries, which are preserved by gradient flow but broken by finite-stepsize stochastic gradient descent (SGD). A canonical example of such a symmetry is scale in homogeneous networks: one can scale up the parameters in one layer and down in the next without changing the network output. Previous work has documented cases in which SGD breaks this symmetry in favor of balancing gradient noise or minimizing fluctuations. Here, we show that the solution geometry of undercomplete linear autoencoders instead selects a preferred sign for scale drift: on the PCA solution manifold, SGD favors large decoder weights. This directed scale drift occurs on a slow timescale, and its dynamics admit an analytically-tractable effective description. However, it cannot continue indefinitely: increasing scale eventually drives the dynamics towards a finite-stepsize stability boundary. The resulting solutions are sharper than a balanced baseline in the sense of the maximum eigenvalue of the loss Hessian, but different sharpness measures can move in opposing directions. Thus, undercomplete autoencoders give a concrete illustration of how loss geometry can convert residual gradient noise into directed motion along a manifold of functionally-equivalent solutions.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno).
Categorie arXiv: Machine Learning (cs.LG) ; Neurons and Cognition (q-bio.NC); Machine Learning (stat.ML).
Pagina del lavoro: https://arxiv.org/abs/2610.03640 — 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 2026-10-05T23:01:38.712Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).