Metraipolis › archive › dc_d2c3cf6ee795
Parameter symmetries determine representational geometry in overparameterized nonlinear networks — arXiv:2609.39078
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.39078
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-07T13:45:47.749Z
- SHA-256 of the archived text
7b1470758fccea4c03bccfb010ef9155b6988f16b7a7b5e6edc73f306c4d0ecc- Stable URL of this page
https://metraipolis.com/documento/dc_d2c3cf6ee795
How to cite
Parameter symmetries determine representational geometry in overparameterized nonlinear networks — arXiv:2609.39078 — arXiv (en), https://arxiv.org/abs/2609.39078 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_d2c3cf6ee795 (Metraipolis archive dc_d2c3cf6ee795, sha256 7b1470758fccea4c03bccfb010ef9155b6988f16b7a7b5e6edc73f306c4d0ecc)
BibTeX
@misc{dcd2c3cf6ee795,
title = {Parameter symmetries determine representational geometry in overparameterized nonlinear networks — arXiv:2609.39078},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.39078},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_d2c3cf6ee795; sha256 7b1470758fccea4c03bccfb010ef9155b6988f16b7a7b5e6edc73f306c4d0ecc},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.39078. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Representations are routinely used across machine learning, psychology, and neuroscience to draw inferences about the computations of biological and artificial systems. Such inferences presume a meaningful link between representational geometry and the computation being performed. For artificial neural networks, however, the extent to which function constrains representation remains unclear. One key obstacle is that these networks admit parameter symmetries: changes in parameterization that preserve function exactly while reshaping representational geometry. Here, we show that a broad class of parameter symmetries acts on representations through just three primitive feature transformations: addition, duplication, and scaling. This feature-level characterization yields a closed-form decomposition of representational geometry into essential and auxiliary components, which makes precise how degeneracy in representational geometry can grow with overparameterization even when function is held fixed. Finally, we show that implementation-level selection rules can resolve this degeneracy, yielding identifiable geometries in which features are weighted according to their contributions to the network's function. Together, our results delineate when representations can support inferences about computation, and when they cannot. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-10-01 in cs.LG). Categorie arXiv: Machine Learning (cs.LG). Pagina del lavoro: https://arxiv.org/abs/2609.39078 — 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 2026-10-07T13:18:10.314Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-10-01?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).