arXiv subjects

J. F. Seabra

Publications and source records attributed to J. F. Seabra.

6 recordsLinked to original sources

arXiv2026

Solving stiff dark matter equations via Jacobian Normalization with Physics-Informed Neural Networks

Stiff differential equations pose a major challenge for Physics-Informed Neural Networks (PINNs), often causing poor convergence. We propose a simple, hyperparameter-free method to address stiffness by normalizing loss residuals with the Jacobian. We provide theoretical indications that Jacobian-based normalization can improve gradient descent and validate it on benchmark stiff ordinary differential equations. We then apply it to a realistic system: the stiff Boltzmann equations (BEs) governing weakly interacting massive particle (WIMP) dark matter (DM). Our approach achieves higher accuracy than attention mechanisms previously proposed for handling stiffness, recovering the full solution where prior methods fail. This is further demonstrated in an inverse problem with a single experimental data point - the observed DM relic density - where our inverse PINNs correctly infer the cross section that solves the BEs in both Standard and alternative cosmologies.

hep-ph↗

arXiv2025

Unraveling particle dark matter with Physics-Informed Neural Networks

We parametrically solve the Boltzmann equations governing freeze-in dark matter (DM) in alternative cosmologies with Physics-Informed Neural Networks (PINNs), a mesh-free method. Through inverse PINNs, using a single DM experimental point -- observed relic density -- we determine the physical attributes of the theory, namely power-law cosmologies, inspired by braneworld scenarios, and particle interaction cross sections. The expansion of the Universe in such alternative cosmologies has been parameterized through a switch-like function reproducing the Hubble law at later times. Without loss of generality, we model more realistically this transition with a smooth function. We predict a distinct pair-wise relationship between power-law exponent and particle interactions: for a given cosmology with negative (positive) exponent, smaller (larger) cross sections are required to reproduce the data. Lastly, via Bayesian methods, we quantify the epistemic uncertainty of theoretical parameters found in inverse problems.

hep-ph↗

arXiv2024

Gradient Boosting MUST taggers for highly-boosted jets

The MUST (Mass Unspecific Supervised Tagging) method has proven to be successful in implementing generic jet taggers capable of discriminating various signals over a wide range of jet masses. We implement the MUST concept by using eXtreme Gradient Boosting (XGBoost) classifiers instead of neural networks (NNs) as previously done. We build both fully-generic and specific multi-pronged taggers, to identify 2, 3, and/or 4-pronged signals from SM QCD background. We show that XGBoost-based taggers are not only easier to optimize and much faster than those based in NNs, but also show quite similar performance, even when testing with signals not used in training. Therefore, they provide a quite efficient alternative machine-learning implementation for generic jet taggers.

hep-ph↗

arXiv2023

Confronting the 95 GeV excesses within the UN2HDM

We consider the small excesses around 95 GeV found in several searches for a new scalar in $γγ$, $ττ$ and $b \bar b$ final states. Instead of trying to accommodate them all, as is usually done in the literature, in the context of a given Standard Model~(SM) extension, we investigate whether it would be possible that one or two of these excesses correspond to an actual new scalar, while the remaining ones are merely statistical fluctuations. To this end, we use as benchmark model the UN2HDM, a SM extension with one scalar doublet, one scalar singlet, and an extra $\text{U}(1)'$ symmetry, which has been previously studied in the context of multiboson cascade decays. We show that most of the possibilities where the excesses in one or two of these channels disappear in the future can be accommodated by type-I or type-III UN2HDMs.

hep-ph↗

arXiv2022

Multiboson signals in the UN2HDM

We address multiboson production from a heavy $Z'$ resonance in the context of the UN2HDM, a standard model extension with an additional $\text{U}(1)'$ symmetry and an enlarged scalar sector with an extra doublet and a singlet. After taking into account theoretical and experimental constraints on the model, it turns out this type of signals - mostly uncovered by current searches - could be sizeable. We focus on three benchmark scenarios, each of them predicting up to 4000 multiboson events with the LHC Run 2 collected luminosity. Anomaly-detection methods could uncover those signals, if present in data.

hep-ph↗

arXiv2021

Mass Unspecific Supervised Tagging (MUST) for boosted jets

Jet identification tools are crucial for new physics searches at the LHC and at future colliders. We introduce the concept of Mass Unspecific Supervised Tagging (MUST) which relies on considering both jet mass and transverse momentum varying over wide ranges as input variables - together with jet substructure observables - of a multivariate tool. This approach not only provides a single efficient tagger for arbitrary ranges of jet mass and transverse momentum, but also an optimal solution for the mass correlation problem inherent to current taggers. By training neural networks, we build MUST-inspired generic and multi-pronged jet taggers which, when tested with various new physics signals, clearly outperform the variables commonly used by experiments to discriminate signal from background. These taggers are also efficient to spot signals for which they have not been trained. Taggers can also be built to determine, with a high degree of confidence, the prongness of a jet, which would be of utmost importance in case a new physics signal is discovered.

hep-ph↗