Tootfinder

Constraining the outputs of ReLU neural networks
Yulia Alexandr, Guido Mont\'ufar
https://arxiv.org/abs/2508.03867 https://arxiv.org/pdf/2508.03867

Constraining the outputs of ReLU neural networks
We introduce a class of algebraic varieties naturally associated with ReLU neural networks, arising from the piecewise linear structure of their outputs across activation regions in input space, and the piecewise multilinear structure in parameter space. By analyzing the rank constraints on the network outputs within each activation region, we derive polynomial equations that characterize the functions representable by the network. We further investigate conditions under which these varieties a…

🇺🇦 #NowPlaying on KEXP's #AstralPlane
Shintaro Sakamoto:
🎵 My Memories Fade
#ShintaroSakamoto
https://shintarosakamotoofficial.bandcamp.com/track/my-memories-fade
https://open.spotify.com/track/2ku6k4TUFANDFT6p5TOnuD

Fishermen in Maceió, Brazil, found to have higher-than-average mercury levels #Brazil …

Fishermen in Maceió, Brazil, found to have higher-than-average mercury levels
Researchers from the Federal University of Alagoas (UFAL) and the State University of Campinas (UNICAMP) in Brazil have found high levels of mercury in the blood and urine of people living on the shores of the Mundaú Lagoon in Maceió, the capital of the Brazilian state of Alagoas. These levels are higher than those found in other populations within the same municipality that have a similar socioeconomic level but live far from the lagoon.

E o projeto Gaia Nefelibata - #Filosofia em Território Virtual - segue no ar!
Saiu o terceiro vídeo do módulo sobre #Bergson - Subjetividade e Metafísica, sob a fala do Prof. Fernando Monegalha (UFAL):

Understanding Learning Invariance in Deep Linear Networks
Hao Duan, Guido Mont\'ufar
https://arxiv.org/abs/2506.13714 https://arx…

Understanding Learning Invariance in Deep Linear Networks
Equivariant and invariant machine learning models exploit symmetries and structural patterns in data to improve sample efficiency. While empirical studies suggest that data-driven methods such as regularization and data augmentation can perform comparably to explicitly invariant models, theoretical insights remain scarce. In this paper, we provide a theoretical comparison of three approaches for achieving invariance: data augmentation, regularization, and hard-wiring. We focus on mean squared e…

🇺🇦 #NowPlaying on KEXP's #PacificNotions
Eli Winter:
🎵 All the Diamonds
#EliWinter
https://tompkinssquare.bandcamp.com/track/all-the-diamonds-2
https://open.spotify.com/track/6R0CXhWuFAwu0KnR7ju7Aj

Replaced article(s) found for nlin.SI. https://arxiv.org/list/nlin.SI/new
[1/1]:
- Effective theory for strongly attractive one-dimensional fermions
Backert, Brauneis, \v{C}ufar, Brand, Hammer, Volosniev