Metraipolis › archive › dc_c68f0cd91f55

Specific Algorithmic Interpretability of Neural Networks: A Case Study on Textures — arXiv:2610.04413

Source
arXiv
Original
https://arxiv.org/abs/2610.04413
License (as archived)
CC0-1.0
Archived
2026-10-06T05:26:12.381Z
SHA-256 of the archived text
6b04a850ac9a839539aea622f3d439ef85d5820209ad65984cd7db38b7d60d1c
Stable URL of this page
https://metraipolis.com/documento/dc_c68f0cd91f55

How to cite

Specific Algorithmic Interpretability of Neural Networks: A Case Study on Textures — arXiv:2610.04413 — arXiv, https://arxiv.org/abs/2610.04413 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_c68f0cd91f55 (Metraipolis archive dc_c68f0cd91f55, sha256 6b04a850ac9a839539aea622f3d439ef85d5820209ad65984cd7db38b7d60d1c)
BibTeX
@misc{dcc68f0cd91f55,
  title = {Specific Algorithmic Interpretability of Neural Networks: A Case Study on Textures — arXiv:2610.04413},
  howpublished = {arXiv},
  url = {https://arxiv.org/abs/2610.04413},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_c68f0cd91f55; sha256 6b04a850ac9a839539aea622f3d439ef85d5820209ad65984cd7db38b7d60d1c},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv, https://arxiv.org/abs/2610.04413. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

We develop a principled framework for constructing neural networks whose specific parameter realizations admit an explicit algorithmic interpretation. Existing algorithm-inspired architectures can explain the computational structure of a network, yet after standard training the learned parameters need not retain a clear relation to the motivating algorithm. We address this gap as follows. First, we model each data point as a sample of a class-dependent stochastic process and assume that statistics of this process can be estimated from a single sample and these statistics are sufficient to distinguish the classes. We then construct a neural network whose initial parameters exactly implement an algorithm for estimating these statistics, making the network fully interpretable. To account for mismatch between the idealized model and real data, we fine-tune this network while controlling its deviation from the algorithmic initialization. The trained network hence roughly retains the interpretation of the initial network. A PAC-Bayesian analysis yields a uniform generalization bound whose complexity term scales with the fine-tuning radius, providing a statistical motivation for our approach. We instantiate the framework for texture classification using the scattering transform to estimate the discriminative statistics.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Machine Learning (cs.LG) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2610.04413 — 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 2026-10-06T05:24:05.850Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).