| Home | << 1 2 3 4 >> |
|
Barenboim, G., Del Debbio, L., Hirn, J., & Sanz, V. (2024). Exploring how a generative AI interprets music. Neural Comput. Appl., 36, 17007–17022.
Abstract: We aim to investigate how closely neural networks (NNs) mimic human thinking. As a step in this direction, we study the behavior of artificial neuron(s) that fire most when the input data score high on some specific emergent concepts. In this paper, we focus on music, where the emergent concepts are those of rhythm, pitch and melody as commonly used by humans. As a black box to pry open, we focus on Google’s MusicVAE, a pre-trained NN that handles music tracks by encoding them in terms of 512 latent variables. We show that several hundreds of these latent variables are “irrelevant” in the sense that can be set to zero with minimal impact on the reconstruction accuracy. The remaining few dozens of latent variables can be sorted by order of relevance by comparing their variance. We show that the first few most relevant variables, and only those, correlate highly with dozens of human-defined measures that describe rhythm and pitch in music pieces, thereby efficiently encapsulating many of these human-understandable concepts in a few nonlinear variables.
|
|
Garcia Navarro, J. E., Fernandez-Prieto, L. M., Villaseñor, A., Sanz, V., Ammirati, J. B., Diaz Suarez, E. A., et al. (2022). Performance of Deep Learning Pickers in Routine Network Processing Applications. Seismol. Res. Lett., 93, 2529–2542.
Abstract: Picking arrival times of P and S phases is a fundamental and time‐consuming task for the routine processing of seismic data acquired by permanent and temporary networks. A large number of automatic pickers have been developed, but to perform well they often require the tuning of multiple parameters to adapt them to each dataset. Despite the great advance in techniques, some problems remain, such as the difficulty to accurately pick S waves and earthquake recordings with a low signal‐to‐noise ratio. Recently, phase pickers based on deep learning (DL) have shown great potential for event identification and arrival‐time picking. However, the general adoption of these methods for the routine processing of monitoring networks has been held back by factors such as the availability of well‐documented software, computational resources, and a gap in knowledge of these methods. In this study, we evaluate recent available DL pickers for earthquake data, comparing the performance of several neural network architectures. We test the selected pickers using three datasets with different characteristics. We found that the analyzed DL pickers (generalized phase detection, PhaseNet, and EQTransformer) perform well in the three tested cases. They are very efficient at ignoring large‐amplitude transient noise and at picking S waves, a task that is often difficult even for experienced analysts. Nevertheless, the performance of the analyzed DL pickers varies widely in terms of sensitivity and false discovery rate, with some pickers missing a significant percentage of true picks and others producing a large number of false positives. There are also variations in run time between DL pickers, with some of them requiring significant resources to process large datasets. In spite of these drawbacks, we show that DL pickers can be used efficiently to process large seismic datasets and obtain results comparable or better than current standard procedures.
|
|
Conde, D., Castillo, F. L., Escobar, C., García, C., Garcia Navarro, J. E., Sanz, V., et al. (2023). Forecasting Geomagnetic Storm Disturbances and Their Uncertainties Using Deep Learning. Space Weather, 21(11), e2023SW003474–27pp.
Abstract: Severe space weather produced by disturbed conditions on the Sun results in harmful effects both for humans in space and in high-latitude flights, and for technological systems such as spacecraft or communications. Also, geomagnetically induced currents (GICs) flowing on long ground-based conductors, such as power networks, potentially threaten critical infrastructures on Earth. The first step in developing an alarm system against GICs is to forecast them. This is a challenging task given the highly non-linear dependencies of the response of the magnetosphere to these perturbations. In the last few years, modern machine-learning models have shown to be very good at predicting magnetic activity indices. However, such complex models are on the one hand difficult to tune, and on the other hand they are known to bring along potentially large prediction uncertainties which are generally difficult to estimate. In this work we aim at predicting the SYM-H index characterizing geomagnetic storms multiple-hour ahead, using public interplanetary magnetic field (IMF) data from the Sun-Earth L1 Lagrange point and SYM-H data. We implement a type of machine-learning model called long short-term memory (LSTM) network. Our scope is to estimate the prediction uncertainties coming from a deep-learning model in the context of forecasting the SYM-H index. These uncertainties will be essential to set reliable alarm thresholds. The resulting uncertainties turn out to be sizable at the critical stages of the geomagnetic storms. Our methodology includes as well an efficient optimization of important hyper-parameters of the LSTM network and robustness tests.
|
|
Sanz, V. (2026). Learning Symmetries in Datasets. Appl. Sci.-Basel, 16(4), 1930–19pp.
Abstract: We investigate how symmetries present in datasets affect the structure of the latent space learned by Variational Autoencoders (VAEs). Understanding symmetries in data is essential because symmetries determine the true degrees of freedom, constrain generalization, and provide physically interpretable coordinates. We therefore study whether a standard, non-equivariant VAE can reveal symmetry-induced dimensional reduction directly from data, without imposing the symmetry in the architecture. By training VAEs on data originating from simple mechanical systems and particle collisions, we analyze the organization of the latent space through a relevance measure that identifies the most meaningful latent directions. We show that when symmetries or approximate symmetries are present, the VAE self-organizes its latent space, effectively compressing the data along a reduced number of latent variables. This behavior captures the intrinsic dimensionality determined by the symmetry constraints and reveals hidden relations among the features. Furthermore, we provide a theoretical analysis of a simple toy model, demonstrating how, under idealized conditions, the latent space aligns with the symmetry directions of the data manifold. We illustrate these findings with examples ranging from two-dimensional datasets with O(2) symmetry to realistic datasets from electron-positron and proton-proton collisions. Our results highlight the potential of unsupervised generative models to expose underlying structures in data and offer a novel approach to symmetry discovery without explicit supervision.
Keywords: generative AI; symmetries; variational autoencoders
|
|
Hirn, J., Garcia, J. E., Montesinos-Navarro, A., Sanchez-Martin, R., Sanz, V., & Verdu, M. (2022). A deep Generative Artificial Intelligence system to predict species coexistence patterns. Methods Ecol. Evol., 13, 1052–1061.
Abstract: Predicting coexistence patterns is a current challenge to understand diversity maintenance, especially in rich communities where these patterns' complexity is magnified through indirect interactions that prevent their approximation with classical experimental approaches. We explore cutting-edge Machine Learning techniques called Generative Artificial Intelligence (GenAI) to predict species coexistence patterns in vegetation patches, training generative adversarial networks (GAN) and variational AutoEncoders (VAE) that are then used to unravel some of the mechanisms behind community assemblage. The GAN accurately reproduces real patches' species composition and plant species' affinity to different soil types, and the VAE also reaches a high level of accuracy, above 99%. Using the artificially generated patches, we found that high-order interactions tend to suppress the positive effects of low-order interactions. Finally, by reconstructing successional trajectories, we could identify the pioneer species with larger potential to generate a high diversity of distinct patches in terms of species composition. Understanding the complexity of species coexistence patterns in diverse ecological communities requires new approaches beyond heuristic rules. Generative Artificial Intelligence can be a powerful tool to this end as it allows to overcome the inherent dimensionality of this challenge.
|
|
Folgado, M. G., & Sanz, V. (2022). Exploring the political pulse of a country using data science tools. J. Comput. Soc. Sci., 5, 987–1000.
Abstract: In this paper we illustrate the use of Data Science techniques to analyse complex human communication. In particular, we consider tweets from leaders of political parties as a dynamical proxy to political programmes and ideas. We also study the temporal evolution of their contents as a reaction to specific events. We analyse levels of positive and negative sentiment in the tweets using new tools adapted to social media. We also train a Fully-Connected Neural Network (FCNN) to recognise the political affiliation of a tweet. The FCNN is able to predict the origin of the tweet with a precision in the range of 71-75%, and the political leaning (left or right) with a precision of around 90%. This study is meant to be viewed as an example of how to use Twitter data and different types of Data Science tools for a political analysis.
|
|
Escudero, M., Rius, N., & Sanz, V. (2017). Sterile neutrino portal to Dark Matter II: exact dark symmetry. Eur. Phys. J. C, 77(6), 397–11pp.
Abstract: We analyze a simple extension of the standard model (SM) with a dark sector composed of a scalar and a fermion, both singlets under the SM gauge group but charged under a dark sector symmetry group. Sterile neutrinos, which are singlets under both groups, mediate the interactions between the dark sector and the SM particles, and generate masses for the active neutrinos via the seesaw mechanism. We explore the parameter space region where the observed Dark Matter relic abundance is determined by the annihilation into sterile neutrinos, both for fermion and scalar Dark Matter particles. The scalar Dark Matter case provides an interesting alternative to the usual Higgs portal scenario. We also study the constraints from direct Dark Matter searches and the prospects for indirect detection via sterile neutrino decays to leptons, which may be able to rule out Dark Matter masses below and around 100 GeV.
|
|
Bagnaschi, E., Ellis, J., Madigan, M., Mimasu, K., Sanz, V., & You, T. (2022). SMEFT analysis of m(W). J. High Energy Phys., 08(8), 308–22pp.
Abstract: We use the Fitmaker tool to incorporate the recent CDF measurement of mW in a global fit to electroweak, Higgs, and diboson data in the Standard Model Effective Field Theory (SMEFT) including dimension-6 operators at linear order. We find that including any one of the SMEFT operators O-HWB, O-HD, O (l) (l) or O ((3)) (H l) with a non-zero coefficient could provide a better fit than the Standard Model, with the strongest pull for O-HD and no tension with other electroweak precision data. We then analyse which tree-level single-field extensions of the Standard Model could generate such operator coefficients with the appropriate sign, and discuss the masses and couplings of these fields that best fit the CDF measurement and other data. In particular, the global fit favours either a singlet Z 0 vector boson, a scalar electroweak triplet with zero hypercharge, or a vector electroweak triplet with unit hypercharge, followed by a singlet heavy neutral lepton, all with masses in the multi-TeV range for unit coupling.
Keywords: Electroweak Precision Physics; SMEFT
|
|
Ellis, J., Madigan, M., Mimasu, K., Sanz, V., & You, T. (2021). Top, Higgs, diboson and electroweak fit to the Standard Model effective field theory. J. High Energy Phys., 04(4), 279–78pp.
Abstract: The search for effective field theory deformations of the Standard Model (SM) is a major goal of particle physics that can benefit from a global approach in the framework of the Standard Model Effective Field Theory (SMEFT). For the first time, we include LHC data on top production and differential distributions together with Higgs production and decay rates and Simplified Template Cross-Section (STXS) measurements in a global fit, as well as precision electroweak and diboson measurements from LEP and the LHC, in a global analysis with SMEFT operators of dimension 6 included linearly. We present the constraints on the coefficients of these operators, both individually and when marginalised, in flavour-universal and top-specific scenarios, studying the interplay of these datasets and the correlations they induce in the SMEFT. We then explore the constraints that our linear SMEFT analysis imposes on specific ultra-violet completions of the Standard Model, including those with single additional fields and low-mass stop squarks. We also present a model-independent search for deformations of the SM that contribute to between two and five SMEFT operator coefficients. In no case do we find any significant evidence for physics beyond the SM. Our underlying Fitmaker public code provides a framework for future generalisations of our analysis, including a quadratic treatment of dimension-6 operators.
|
|
Cepedello, R., Esser, F., Hirsch, M., & Sanz, V. (2024). Fermionic UV models for neutral triple gauge boson vertices. J. High Energy Phys., 07(7), 275–28pp.
Abstract: Searches for anomalous neutral triple gauge boson couplings (NTGCs) provide important tests for the gauge structure of the standard model. In SMEFT (“standard model effective field theory”) NTGCs appear only at the level of dimension-8 operators. While the phenomenology of these operators has been discussed extensively in the literature, renormalizable UV models that can generate these operators are scarce. In this work, we study a variety of extensions of the SM with heavy fermions and calculate their matching to d = 8 NTGC operators. We point out that the complete matching of UV models requires four different CP-conserving d = 8 operators and that the single CPC d = 8 operator, most commonly used by the experimental collaborations, does not describe all possible NTGC form factors. Despite stringent experimental constraints on NTGCs, limits on the scale of UV models are relatively weak, because their contributions are doubly suppressed (being d = 8 and 1-loop). We suggest a series of benchmark UV scenarios suitable for interpreting searches for NTGCs in the upcoming LHC runs, obtain their current limits and provide estimates for the expected sensitivity of the high-luminosity LHC.
|