Metraipolis › archive › dc_7cb97f2103a0

Not all solutions are created equal: An analytical dissociation of functional and representational similarity in deep linear neural networks — arXiv:2609.38998

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.38998
License (as archived)
CC0-1.0
Archived
2026-10-07T13:45:43.797Z
SHA-256 of the archived text
262c3ed0be21aeec2f027c23dde08eb83e59886f0c9897e9f4e4f5b549f53c74
Stable URL of this page
https://metraipolis.com/documento/dc_7cb97f2103a0

How to cite

Not all solutions are created equal: An analytical dissociation of functional and representational similarity in deep linear neural networks — arXiv:2609.38998 — arXiv (en), https://arxiv.org/abs/2609.38998 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_7cb97f2103a0 (Metraipolis archive dc_7cb97f2103a0, sha256 262c3ed0be21aeec2f027c23dde08eb83e59886f0c9897e9f4e4f5b549f53c74)
BibTeX
@misc{dc7cb97f2103a0,
  title = {Not all solutions are created equal: An analytical dissociation of functional and representational similarity in deep linear neural networks — arXiv:2609.38998},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.38998},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_7cb97f2103a0; sha256 262c3ed0be21aeec2f027c23dde08eb83e59886f0c9897e9f4e4f5b549f53c74},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.38998. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

A foundational principle of connectionism is that perception, action, and cognition emerge from parallel computations among simple, interconnected units that generate and rely on neural representations. Accordingly, researchers employ multivariate pattern analysis to decode and compare the neural codes of artificial and biological networks, aiming to uncover their functions. However, there is limited analytical understanding of how a network's representation and function relate, despite this being essential to any quantitative notion of underlying function or functional similarity. We address this question using analysable two-layer linear networks and numerical simulations in non-linear networks. We find that function and representation are dissociated, allowing representational similarity without functional similarity and vice versa. Further, we show that neither robustness to input noise nor the level of generalization error constrain representations to the task. In contrast, networks robust to parameter noise have limited representational flexibility and must employ task-specific representations. Our findings suggest that representational alignment reflects computational advantages beyond functional alignment alone, with significant implications for interpreting and comparing the representations of connectionist systems.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-10-01 in cs.LG).
Categorie arXiv: Machine Learning (cs.LG) ; Neurons and Cognition (q-bio.NC).
Pagina del lavoro: https://arxiv.org/abs/2609.38998 — 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 2026-10-07T13:18:10.303Z 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).