Metraipolis › archive › dc_34dbb7f74ca8
A Counterexample to Fourier Alignment in Single-Neuron Modular Addition — arXiv:2608.04451
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2608.04451
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-31T11:35:41.261Z
- SHA-256 of the archived text
29d68e109741a965e9c25b09646f6bb2295863cac828713f63746bf68140b21a- Stable URL of this page
https://metraipolis.com/documento/dc_34dbb7f74ca8
How to cite
A Counterexample to Fourier Alignment in Single-Neuron Modular Addition — arXiv:2608.04451 — arXiv (en), https://arxiv.org/abs/2608.04451 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_34dbb7f74ca8 (Metraipolis archive dc_34dbb7f74ca8, sha256 29d68e109741a965e9c25b09646f6bb2295863cac828713f63746bf68140b21a)
BibTeX
@misc{dc34dbb7f74ca8,
title = {A Counterexample to Fourier Alignment in Single-Neuron Modular Addition — arXiv:2608.04451},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2608.04451},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_34dbb7f74ca8; sha256 29d68e109741a965e9c25b09646f6bb2295863cac828713f63746bf68140b21a},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.04451. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
We give a negative solution to MAIS-O60. We first construct an example in which an initially active ReLU neuron becomes completely inactive in finite time and thereafter remains frozen at a limit whose Fourier energy is equally distributed among all nonzero real frequency classes. The counterexample holds on an open set of initial conditions and therefore occurs with positive probability under Gaussian initialization. An appendix prepared by GPT-5.6 Sol strengthens the counterexample by showing that the same failure can occur for every Clarke trajectory from an open set of initial conditions, under the convention $\mathrm{ReLU}'(0)=0$, for smooth dead-zone approximations of ReLU, and for fixed-step full-batch gradient descent. Thus, single-frequency alignment is not a general consequence of training a single neuron on modular addition.
Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-06 in cs.LG).
Categorie arXiv: Optimization and Control (math.OC) ; Machine Learning (cs.LG); Neural and Evolutionary Computing (cs.NE).
Pagina del lavoro: https://arxiv.org/abs/2608.04451 — 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 2026-10-31T11:24:48.830Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-08-06?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).