Metraipolis › archive › dc_600ee27f2a6f

$β$-VAEs as Effective Theories: Tolerance-Dependent Dimension — arXiv:2608.10599

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.10599
License (as archived)
CC0-1.0
Archived
2026-10-19T06:19:23.892Z
SHA-256 of the archived text
2962fa5ca4ad809ba01d51340810a775638eead0089ef093adabcba0d1ebb782
Stable URL of this page
https://metraipolis.com/documento/dc_600ee27f2a6f

How to cite

$β$-VAEs as Effective Theories: Tolerance-Dependent Dimension — arXiv:2608.10599 — arXiv (en), https://arxiv.org/abs/2608.10599 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_600ee27f2a6f (Metraipolis archive dc_600ee27f2a6f, sha256 2962fa5ca4ad809ba01d51340810a775638eead0089ef093adabcba0d1ebb782)
BibTeX
@misc{dc600ee27f2a6f,
  title = {$β$-VAEs as Effective Theories: Tolerance-Dependent Dimension — arXiv:2608.10599},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.10599},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_600ee27f2a6f; sha256 2962fa5ca4ad809ba01d51340810a775638eead0089ef093adabcba0d1ebb782},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.10599. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

In a $\beta$-VAE, increasing the regularization strength acts as a spectral cutoff by collapsing low-utility latent coordinates. In the linear Gaussian VAE, the collapse order matches the ranking of reconstruction utilities exactly, because both are set by the PCA spectrum. We ask which parts of this picture survive in fully connected nonlinear VAEs trained on WorldClim. We find that nonlinear interactions shift and broaden collapse onsets, so thresholds no longer coincide exactly with utilities. However, the common ordering is preserved over the resolved ranks, so the spectral cutoff still acts as a utility cutoff and the effective-description logic carries through. The resulting effective-dimension curves reveal a head--tail tradeoff: increasing depth concentrates utility into the first few coordinates but worsens tail fidelity.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-12 in cs.LG).
Categorie arXiv: Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2608.10599 — 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 2026-10-19T06:11:45.087Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-08-12?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).