arXiv subjects

Dário Passos

Publications and source records attributed to Dário Passos.

5 recordsLinked to original sources

arXiv2026

Using Data-Derived Priors to Guide CNN Architecture Design for NIR Chemometrics

Convolutional neural networks (CNN) for near-infrared (NIR) chemometrics are often designed using generic architectural rules, although spectral datasets differ in sampling, smoothness, redundancy, and sample size. We tested whether these properties can provide empirical priors for CNN design. Across 25 NIR regression tasks, we computed descriptors of dataset size, spectral length and spacing, entropy, intrinsic rank, autocorrelation, and wavelet-scale structure. Two interpretable 1D-CNN scaffolds (a minimal single-convolution model and an extended shallow model with optional branching, dilation, etc) were optimized using five-fold cross-validated Bayesian hyperparameter optimization (HPO). Relationships extracted from near-optimal trials were converted into warm-start heuristics and evaluated directly and through leave-one-dataset-out (LODO) validation. The clearest relationships involved convolutional receptive fields. In the minimal CNN, the preferred kernel fraction decreased with spectral entropy and intrinsic rank, increased with the wavelet energy-support fraction, and the learning rate tended to decrease with training-set size. Direct and LODO heuristics were competitive with HPO, with median test-RMSE ratios of 0.953 and 1.017, respectively. The extended CNN showed similar but less transferable structure across branch usage, dilation, dropout, filter counts, and receptive-field choices. Ten stochastic refits showed seed sensitivity comparable to that of HPO-selected configurations. In a separate experiment, joint preprocessing and CNN HPO outperformed standardized-spectra HPO in 19 of 25 tasks, although gains were dataset-dependent. These results show that spectral descriptors can provide practical CNN design priors, guiding shallow NIR models toward plausible hyperparameter regions before target-specific tuning

cs.LG↗

arXiv2026

Convolutional Neural Networks in Vis-NIR Chemometrics: From Contradiction to Conditional Design

Near-infrared (NIR and Vis-NIR) spectroscopy is widely used for rapid, non-destructive analysis in food, agriculture, pharmaceuticals, process analytical technology, and bioprocess monitoring. Nevertheless, deep-learning studies in NIR chemometrics often reach conflicting conclusions about convolutional neural network (CNN) design, including kernel size, depth, preprocessing, model complexity, and transfer robustness. This review argues that many apparent contradictions reflect incomplete experimental conditioning rather than incompatible findings. CNN performance depends on interactions among spectral physics, dataset regime, acquisition protocol, validation design, and deployment conditions. We organize the literature around three moderators. First, NIR signals are indirect, highly collinear, and shaped by broad overlapping bands, scattering, temperature, and matrix effects. Second, CNN design should be interpreted through receptive-field reasoning: kernel size, depth, dilation, and multi-scale branches determine the wavelength span available to the model, whereas the effective receptive field indicates which parts are actually used. Third, validation design can behave as a hidden hyperparameter because random splits may reward architectures that exploit shared batch, instrument, season, or process-run structure instead of transferable chemical information. We therefore propose a conditional design framework in which preprocessing, architecture, hyperparameter optimization, transfer evaluation, interpretability, and reproducibility are treated as coupled components. Rather than seeking a universally optimal CNN, the framework aims to support physics-aware, shift-aware, and reproducible model comparison in NIR chemometrics.

cs.LG↗

arXiv2014

Oscillator models of the solar cycle: Towards the development of inversion methods

This article reviews some of the leading results obtained in solar dynamo physics by using temporal oscillator models as a tool to interpret observational data and dynamo model predictions. We discuss how solar observational data such as the sunspot number is used to infer the leading quantities responsible for the solar variability during the last few centuries. Moreover, we discuss the advantages and difficulties of using inversion methods (or backward methods) over forward methods to interpret the solar dynamo data. We argue that this approach could help us to have a better insight about the leading physical processes responsible for solar dynamo, in a similar manner as helioseismology has helped to achieve a better insight on the thermodynamic structure and flow dynamics in the Sun's interior.

astro-ph.SR↗

arXiv2014

A Stochastically Forced Time Delay Solar Dynamo Model: Self-Consistent Recovery from a Maunder-like Grand Minimum Necessitates a Mean-Field Alpha Effect

Fluctuations in the Sun's magnetic activity, including episodes of grand minima such as the Maunder minimum have important consequences for space and planetary environments. However, the underlying dynamics of such extreme fluctuations remain ill-understood. Here we use a novel mathematical model based on stochastically forced, non-linear delay differential equations to study solar cycle fluctuations, in which, time delays capture the physics of magnetic flux transport between spatially segregated dynamo source regions in the solar interior. Using this model we explicitly demonstrate that the Babcock-Leighton poloidal field source based on dispersal of tilted bipolar sunspot flux, alone, can not recover the sunspot cycle from a grand minimum. We find that an additional poloidal field source effective on weak fields--the mean-field alpha-effect driven by helical turbulence--is necessary for self-consistent recovery of the sunspot cycle from grand minima episodes.

astro-ph.SR↗

arXiv2012

Effects of cyclic fluctuations in meridional circulation using a low order dynamo model

We develop and subsequently explore the solution space of a simple flux transport dynamo model that incorporates a time dependent large scale meridional circulation. Based on recent observations we prescribed an analytical form for the amplitude of this circulation and study its impact in the evolution of the magnetic field. We find that cyclic variations in the amplitude and frequency of the meridional flow affect the strength of the solar cycle. Variations in the amplitude of the fluctuations influence the shape of the solar cycle but are only relevant to the cycle's strength variations when they occur at a frequency different from or out of phase of the solar cycle's.

astro-ph.SR↗