Caron, S., Ruiz de Austri, R., & Zhang, Z. Y. (2023). Mixture-of-Theories training: can we find new physics and anomalies better by mixing physical theories? J. High Energy Phys., 03(3), 004–37pp.
Abstract: Model-independent search strategies have been increasingly proposed in recent years because on the one hand there has been no clear signal for new physics and on the other hand there is a lack of a highly probable and parameter-free extension of the standard model. For these reasons, there is no simple search target so far. In this work, we try to take a new direction and ask the question: bearing in mind that we have a large number of new physics theories that go beyond the Standard Model and may contain a grain of truth, can we improve our search strategy for unknown signals by using them “in combination”? In particular, we show that a signal hypothesis based on a large, intermingled set of many different theoretical signal models can be a superior approach to find an unknown BSM signal. Applied to a recent data challenge, we show that “mixture-of-theories training” outperforms strategies that optimize signal regions with a single BSM model as well as most unsupervised strategies. Applications of this work include anomaly detection and the definition of signal regions in the search for signals of new physics.
|
Gomez Ambrosio, R., ter Hoeve, J., Madigan, M., Rojo, J., & Sanz, V. (2023). Unbinned multivariate observables for global SMEFT analyses from machine learning. J. High Energy Phys., 03(3), 033–66pp.
Abstract: Theoretical interpretations of particle physics data, such as the determination of the Wilson coefficients of the Standard Model Effective Field Theory (SMEFT), often involve the inference of multiple parameters from a global dataset. Optimizing such interpretations requires the identification of observables that exhibit the highest possible sensitivity to the underlying theory parameters. In this work we develop a flexible open source frame-work, ML4EFT, enabling the integration of unbinned multivariate observables into global SMEFT fits. As compared to traditional measurements, such observables enhance the sensitivity to the theory parameters by preventing the information loss incurred when binning in a subset of final-state kinematic variables. Our strategy combines machine learning regression and classification techniques to parameterize high-dimensional likelihood ratios, using the Monte Carlo replica method to estimate and propagate methodological uncertainties. As a proof of concept we construct unbinned multivariate observables for top-quark pair and Higgs+Z production at the LHC, demonstrate their impact on the SMEFT parameter space as compared to binned measurements, and study the improved constraints associated to multivariate inputs. Since the number of neural networks to be trained scales quadratically with the number of parameters and can be fully parallelized, the ML4EFT framework is well-suited to construct unbinned multivariate observables which depend on up to tens of EFT coefficients, as required in global fits.
|
De Romeri, V., Miranda, O. G., Papoulias, D. K., Sanchez Garcia, G., Tortola, M., & Valle, J. W. F. (2023). Physics implications of a combined analysis of COHERENT CsI and LAr data. J. High Energy Phys., 04(4), 035–41pp.
Abstract: The observation of coherent elastic neutrino nucleus scattering has opened the window to many physics opportunities. This process has been measured by the COHERENT Collaboration using two different targets, first CsI and then argon. Recently, the COHERENT Collaboration has updated the CsI data analysis with a higher statistics and an improved understanding of systematics. Here we perform a detailed statistical analysis of the full CsI data and combine it with the previous argon result. We discuss a vast array of implications, from tests of the Standard Model to new physics probes. In our analyses we take into account experimental uncertainties associated to the efficiency as well as the timing distribution of neutrino fluxes, making our results rather robust. In particular, we update previous measurements of the weak mixing angle and the neutron root mean square charge radius for CsI and argon. We also update the constraints on new physics scenarios including neutrino nonstandard interactions and the most general case of neutrino generalized interactions, as well as the possibility of light mediators. Finally, constraints on neutrino electromagnetic properties are also examined, including the conversion to sterile neutrino states. In many cases, the inclusion of the recent CsI data leads to a dramatic improvement of bounds.
|
Beltran, R., Cottin, G., Hirsch, M., Titov, A., & Wang, Z. S. (2023). Reinterpretation of searches for long-lived particles from meson decays. J. High Energy Phys., 05(5), 031–31pp.
Abstract: Many models beyond the Standard Model predict light and feebly interacting particles that are often long-lived. These long-lived particles (LLPs) in many cases can be produced from meson decays. In this work, we propose a simple and quick reinterpretation method for models predicting LLPs produced from meson decays. With the method, we are not required to run Monte-Carlo simulation, implement detector geometries and efficiencies, or apply experimental cuts in an event analysis, as typically done in recasting and reinterpretation works. The main ingredients our method requires are only the theoretical input, allowing for computation of the production and decay rates of the LLPs. There are two conditions for the method to work: firstly, the LLPs in the models considered should be produced from a set of mesons with similar mass and lifetime (or the same meson) and second, the LLPs should, in general, have a lab-frame decay length much larger than the distance between the interaction point and the detector. As an example, we use this method to reinterpret exclusion bounds on heavy neutral leptons (HNLs) in the minimal “3+1” scenario, into those for HNLs in the general effective-field-theory framework as well as for axion-like particles. We are able to reproduce existing results, and obtain new bounds via reinterpretation of past experimental results, in particular, from CHARM and Belle.
|
Davier, M., Diaz-Calderon, D., Malaescu, B., Pich, A., Rodriguez-Sanchez, A., & Zhang, Z. (2023). The Euclidean Adler function and its interplay with Delta alpha(had)(QED) and alpha(s). J. High Energy Phys., 04(4), 067–57pp.
Abstract: Three different approaches to precisely describe the Adler function in the Euclidean regime at around 2 GeVs are available: dispersion relations based on the hadronic production data in e(+)e(-) annihilation, lattice simulations and perturbative QCD (pQCD). We make a comprehensive study of the perturbative approach, supplemented with the leading power corrections in the operator product expansion. All known contributions are included, with a careful assessment of uncertainties. The pQCD predictions are compared with the Adler functions extracted from ?a( QED)(had)(Q(2)), using both the DHMZ compilation of e(+)e(-) data and published lattice results. Taking as input the FLAG value of a(s), the pQCD Adler function turns out to be in good agreement with the lattice data, while the dispersive results lie systematically below them. Finally, we explore the sensitivity to a(s) of the direct comparison between the data-driven, lattice and QCD Euclidean Adler functions. The precision with which the renormalisation group equation can be tested is also evaluated.
|
