arXiv subjects

Diego D. Lopes

Publications and source records attributed to Diego D. Lopes.

4 recordsLinked to original sources

arXiv2023

Deep Learning Criminal Networks

Recent advances in deep learning methods have enabled researchers to develop and apply algorithms for the analysis and modeling of complex networks. These advances have sparked a surge of interest at the interface between network science and machine learning. Despite this, the use of machine learning methods to investigate criminal networks remains surprisingly scarce. Here, we explore the potential of graph convolutional networks to learn patterns among networked criminals and to predict various properties of criminal networks. Using empirical data from political corruption, criminal police intelligence, and criminal financial networks, we develop a series of deep learning models based on the GraphSAGE framework that are able to recover missing criminal partnerships, distinguish among types of associations, predict the amount of money exchanged among criminal agents, and even anticipate partnerships and recidivism of criminals during the growth dynamics of corruption networks, all with impressive accuracy. Our deep learning models significantly outperform previous shallow learning approaches and produce high-quality embeddings for node and edge properties. Moreover, these models inherit all the advantages of the GraphSAGE framework, including the generalization to unseen nodes and scaling up to large graph structures.

physics.soc-ph↗

arXiv2023

Tracking axion-like particles at the LHC

Highly boosted axion-like particles decaying into photon pairs are notoriously hard to detect at the LHC. The collimated decay photons cannot be individually reconstructed using only electromagnetic calorimeter information, making the signal less distinguishable from background. In this note we propose a search strategy to address this issue, exploiting the fact that a fraction of the decay photons convert into electron-positron pairs inside the tracking detector. The resulting tracks can be resolved with high resolution, allowing to separate the two collimated photons and resolve a displaced decay vertex. To demonstrate the effectiveness of this approach, we apply it to ALPs in the challenging MeV-GeV range produced via vector boson fusion. We find that such a search could give sensitivity to untested parameter space.

hep-ph↗

arXiv2022

Machine Learning Partners in Criminal Networks

Recent research has shown that criminal networks have complex organizational structures, but whether this can be used to predict static and dynamic properties of criminal networks remains little explored. Here, by combining graph representation learning and machine learning methods, we show that structural properties of political corruption, police intelligence, and money laundering networks can be used to recover missing criminal partnerships, distinguish among different types of criminal and legal associations, as well as predict the total amount of money exchanged among criminal agents, all with outstanding accuracy. We also show that our approach can anticipate future criminal associations during the dynamic growth of corruption networks with significant accuracy. Thus, similar to evidence found at crime scenes, we conclude that structural patterns of criminal networks carry crucial information about illegal activities, which allows machine learning methods to predict missing information and even anticipate future criminal behavior.

physics.soc-ph↗

arXiv2020

Probing ALP-Sterile Neutrino Couplings at the LHC

In this work, prospects to probe an overlooked facet of axion-like particles (ALPs) -- their potential couplings to sterile neutrinos -- are presented. We found that mono-photon searches have the potential to constrain ALP couplings to sterile neutrinos when a new heavy scalar boosts the ALP decay yields. Working within an effective field theory (EFT) approach, we scan the parameters space to establish the reach of the 13 TeV LHC to probe such couplings. We found regions of the parameters space evading several experimental constraints that can be probed at the LHC. Moreover, a complementary role between the LHC and various experiments that search for axions and ALPs can be anticipated for models where ALPs interact with sterile neutrinos. We also present the UV realization of a model having an axion-like particle, a heavy scalar and sterile neutrinos whose parameters are spanned by our EFT approach. The proposed model contains a type of seesaw mechanism for generating masses for the active neutrinos along with sterile neutrinos involving the high energy scale of the spontaneous breaking of the global symmetry associated to the ALP. Some benchmark points of this model can be discovered at the 13 TeV LHC with 300 fb$^{-1}$.

hep-ph↗