Metraipolis › archive › dc_130b6077b008

Does Compression Imply Generalization? A Minimum Description Length Perspective — arXiv:2610.21376

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.21376
License (as archived)
CC0-1.0
Archived
2026-10-16T10:22:29.333Z
SHA-256 of the archived text
4624baca942413df75c39f9659634e6a01ed114a667549953738ec5e888d79ac
Stable URL of this page
https://metraipolis.com/documento/dc_130b6077b008

How to cite

Does Compression Imply Generalization? A Minimum Description Length Perspective — arXiv:2610.21376 — arXiv (en), https://arxiv.org/abs/2610.21376 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_130b6077b008 (Metraipolis archive dc_130b6077b008, sha256 4624baca942413df75c39f9659634e6a01ed114a667549953738ec5e888d79ac)
BibTeX
@misc{dc130b6077b008,
  title = {Does Compression Imply Generalization? A Minimum Description Length Perspective — arXiv:2610.21376},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.21376},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_130b6077b008; sha256 4624baca942413df75c39f9659634e6a01ed114a667549953738ec5e888d79ac},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.21376. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

A central question in modern learning theory is what notion of compression, if any, explains why over-parameterized neural networks generalize. Existing answers measure complexity at different levels, ranging from weights to representations; however, estimating the associated quantities either requires held-out data in principle or relies on the challenging (ultra) high-dimensional mutual information estimation. To overcome these challenges, we approach the question through the minimum description length (MDL) principle, and specifically, we propose the prequential regret as a complementary signal that is computable from the training trajectory alone. We establish that, in the (nearly) interpolating regime, the expected prequential regret upper-bounds the description length of the trained parameter under the algorithm-induced distribution, and that this description length translates into a high-probability generalization bound in the PAC-MDL setting. Along the way, we show that the encoder-complexity correction recently added to the algorithm-level information bottleneck (IB) is upper-bounded by an MDL parameter description length, \new{bringing together} the two views of compression from MDL and IB. Empirically, the prequential regret tracks the generalization error on CIFAR-10 and CIFAR-100, with Pearson correlations exceeding 0.9.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno).
Categorie arXiv: Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2610.21376 — 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 2026-10-16T08:48:02.710Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).