Zobrazeno 1 - 10
of 10 181
pro vyhledávání: '"Novello, A."'
This work investigates the structure and representation capacity of $sinusoidal$ MLPs, which have recently shown promising results in encoding low-dimensional signals. This success can be attributed to its smoothness and high representation capacity.
Externí odkaz:
http://arxiv.org/abs/2407.21121
Autor:
Novello, Mario, Antunes, Vicente
We present solutions corresponding to rotational configurations in the recently proposed Geometric Scalar Gravity (GSG) theory. The solutions obtained here have the important property that the associated closed time-like curves are always restricted
Externí odkaz:
http://arxiv.org/abs/2407.09663
Since the seminal paper of Hendrycks et al. arXiv:1610.02136, Post-hoc deep Out-of-Distribution (OOD) detection has expanded rapidly. As a result, practitioners working on safety-critical applications and seeking to improve the robustness of a neural
Externí odkaz:
http://arxiv.org/abs/2407.07135
Research on Out-Of-Distribution (OOD) detection focuses mainly on building scores that efficiently distinguish OOD data from In Distribution (ID) data. On the other hand, Conformal Prediction (CP) uses non-conformity scores to construct prediction se
Externí odkaz:
http://arxiv.org/abs/2403.11532
Autor:
Novello, Mario, Toniato, Júnior D.
In the general relativity theory the basic ingredient to describe gravity is the geometry, which interacts with all forms of matter and energy, and as such, the metric could be interpreted as a true physical quantity. However the metric is not matter
Externí odkaz:
http://arxiv.org/abs/2402.16163
Autor:
Anjos, Fábio dos, Novello, Mario
The accepted idea that the expansion of the universe is accelerating needs, for compatibility to general relativity, the introduction of some unusual forms of matter. However, several authors have proposed that instead of making weird hypothesis on s
Externí odkaz:
http://arxiv.org/abs/2402.13860
We explore sinusoidal neural networks to represent periodic tileable textures. Our approach leverages the Fourier series by initializing the first layer of a sinusoidal neural network with integer frequencies with a period $P$. We prove that the comp
Externí odkaz:
http://arxiv.org/abs/2402.02208
Autor:
Novello, Nicola, Tonello, Andrea M.
In deep learning, classification tasks are formalized as optimization problems often solved via the minimization of the cross-entropy. However, recent advancements in the design of objective functions allow the usage of the $f$-divergence to generali
Externí odkaz:
http://arxiv.org/abs/2401.01268
Deep neural networks (DNNs) often fail silently with over-confident predictions on out-of-distribution (OOD) samples, posing risks in real-world deployments. Existing techniques predominantly emphasize either the feature representation space or the g
Externí odkaz:
http://arxiv.org/abs/2312.14427
Publikováno v:
Climate of the Past, Vol 20, Pp 2117-2141 (2024)
Paleoclimatological field reconstructions are valuable for understanding past hydroclimatic variability, which is crucial for assessing potential future hydroclimate changes. Despite being as impactful on societies as temperature variability, hydrocl
Externí odkaz:
https://doaj.org/article/aa9a0e3c52ed4e7a9dc19a1aa178a0dc
