Nonlocality Effect in the Tunneling of Alpha Radioactivity with the Aid of Machine Learning
Jinyu Hu 1 and Chen Wu 1 · 2025
Read Nonlocality Effect in the Tunneling of Alpha Radioactivity with the Aid of Machine Learning by Jinyu Hu 1 and Chen Wu 1 (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2510-18199.txt.
Introduction
Jinyu Hu and Chen Wu sharpen predictions of alpha decay by adding a nonlocal effect and machine learning tuning. Alpha particles do not behave as purely local objects inside the nucleus, and folding that fact into a two-potential model already improved results in earlier work. Here three classic learning models optimize the extended approach against reference data.
Decision tree and gradient boosting models agreed best with the data. Their spread dropped by over half against the plain two-potential model without nonlocality. The random forest variant lagged behind. Nuclear theorists get a ranked sense of which simple learners best polish a physics-based decay model.
Catalog record
| CREATOR | Jinyu Hu 1 and Chen Wu 1 |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 29 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2510-18199
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Nonlocality Effect in the Tunneling of Alpha Radioactivity with the Aid of Machine Learning Preprint: APS/123-QED Jinyu Hu 1 and Chen Wu 1 Abstract Recently, building upon the research findings of E. L. Medeiros Medeiros et al. 2022 , we have extended the α \alpha -particle nonlocality effect to the two-potential approach (TPA) Hu and Wu 2025 . This extension demonstrates that the integration of the α \alpha -particle nonlocality effect into TPA yields relatively favorable results. In the present work, we employ machine learning methods to further optimize the aforementioned approach, specifically utilizing three classical machine learning models: decision tree regression, random forest regression, and XGBRegressor. Among these models, both the decision tree regression and XGBRegressor models exhibit the highest degree of agreement with the reference data, whereas the random forest regression model shows inferior performance. In terms of standard deviation, the results derived from the decision tree regression and XGBRegressor models represent improvements of 54.5% and 53.7%, respectively, compared to the TPA that does not account for the coordinate-dependent effective mass of α \alpha particles. Furthermore, we extend the decision tree regression and XGBRegressor models to predict the α \alpha -decay half-lives of 20 even-even nuclei with atomic numbers Z=118 and Z=120. Subsequently, the superheavy nucleus half-life predictions generated by our proposed models are compared with those from two established benchmarks: the improved eight-parameter Deng-Zhang-Royer (DZR) Deng et al. 2020 model and the new empirical expression (denoted as "New+D") Denisov 2024 proposed by V. Yu. Denisov, which explicitly incorporates nuclear deformation effects. Overall, the predictions from these models and formulas are generally consistent. Notably, the predictions of the decision tree regression model show a high level of consistency with those of the New+D expression, while the XGBRegressor model exhibits deviations from the other two comparative models. I INTRODUCTION Since the 20th century, and particularly following the discovery of natural radioactive decay, \alpha -decay has remained a long-standing research focus in nuclear physics. With advances in experimental techniques enabling the laboratory synthesis of superheavy nuclei, interest in the systematic investigation of \alpha -decay processes has grown significantly. This decay mode was first identified as \alpha -radiation emitted by uranium and its compounds; Ernest Rutherford initially characterized it as a process wherein a parent nucleus emits a 4 He nucleus. Subsequently, the underlying mechanism of \alpha -decay was elucidated by quantum tunneling theory, independently proposed by Gurney and Condon Gurney and Condon 1928 and by Gamow Gamow 1928 .Since then, to deepen the understanding of \alpha -disintegration, numerous researchers have explored key nuclear characteristics, including nuclear shape coexistence, low-lying states, shell closure effects, energy level structures, and ground-state properties Matsuse et al. 1975 ; Wauters et al. 1994 ; Kucuk et al. 2020 ; Wang et al. 2017 ; Xiao et al. 2020 . Furthermore, studies of \alpha -decay processes serve as a critical tool for the identification of synthesized heavy and superheavy nuclei. Driven by advances in modern detector technology and the development of realistic nuclear potentials, significant progress has been achieved in both the theoretical Buck et al. 1992 ; Duarte et al. 1998 ; Goncalves and Duarte 1993 ; Basu 2003 ; Chowdhury et al. 2006 and experimental Ma et al. 2020 ; Oganessian et al. 2011 ; Oganessian et al. 2004 ; Zhang et al. 2021 ; Oganessian et al. 2000 ; Oganessian 2011 investigations of \alpha -decay. In the theoretical realm, numerous models and approaches have been proposed to study \alpha -decay, including the two-potential approach Gurvitz and Kalbermann 1987 , fission-like model Buck et al. 1993 , density-dependent M3Y effective interaction Samanta et al. 2007 , cluster-formation model Ahmed 2017 , and generalized liquid drop model Zhang et al. 2006 . Complementing these theoretical frameworks, a wealth of empirical formulas—rooted in the classic Geiger-Nuttall (G-N) law and quantum tunneling effect—have also been developed for this decay mode. Examples include the Viola-Seaborg-Sobiczewski (VSS) formula Viola Jr and Seaborg 1966 , Horoi formula, Universal decay law Qi et al. 2009 , Royer formula Royer 2000 , modified YQZR (MYQZR) formula, Deng-Zhang-Royer (DUR) formula Deng et al. 2020 , and the new empirical expression accounting for nuclear deformation (denoted as "New+D") Denisov 2024 . In terms of experimental progress Hofmann and Münzenberg 2000 ; Hamilton et al. 2013 ; Nazarewicz 2018 ; Giuliani et al. 2019 , the synthesis of superheavy elements in laboratories such as GSI, RIKEN, and JINR primarily relies on cold fusion reactions between 208 Pb or 209 Bi targets and nuclear beams with mass number A>50. As an extension of the periodic table, the superheavy element with atomic number Z=118 was successfully synthesized via 48 Ca-induced hot fusion reactions, using the actinide element californium (Cf) as the target material Oganessian 2011 . Recently, machine learning has attracted significant attention from researchers in physics due to its capability to address complex problems in nonlinear systems. For example, in nuclear physics, it has been applied to predict nuclear masses, \beta -decay half-lives, \alpha -decay half-lives, charge radii, and neutron drip line. Specifically, the Bayesian neural network method has been successfully applied to predict charge radii, \beta -decay half-life, and fission. As a result, it has become a research approach that attracts considerable attention in nuclear physics. In addition, numerous other machine learning methods—such as artificial neural networks and convolutional neural networks—are also actively utilized in nuclear physics research. More importantly, support vector machines have also achieved significant breakthroughs in the field of nuclear physics in recent years. Specifically, when Amir Jalili Jalili et al. 2024 et. al used support vector machines to conduct a detailed calculation of \alpha -decay half-lives, with the root mean square errors of 0.427, which is significantly lower than that of other models for calculating \alpha -decay half-lives. It is precisely due to the diverse successes of machine learning in the field of nuclear physics that we are motivated to employ this approach to assist in predicting \alpha -decay of even-even nuclei. In this paper, we systematically study the nonlocality effect in \alpha -decay of even-even nuclei by using machine learning, specifically using decision tree regression, random forest regression, and XGBRegressor models, to optimize the coordinate-dependent effective mass of mass parameters for alpha particles. What’s more, based on the research by E. L. Medeiros Medeiros et al. 2022 et al, we use machine learning models to optimize the coordinate-dependent effective mass parameters for each individual nucleus. It is found that the results obtained by the decision tree regression model and the XGBRegressor model have improved by 54.5% and 53.7%, respectively, compared with the two-potential approach without the coordinate-dependent effective mass for alpha particles in terms of the standard deviation. As an application, we further extend the present model to predict the \alpha decay half-lives of 20 even-even nuclei with Z=118 and Z=120 , compared to DUR and New+D. Finally, we find the predictions of the present model and New+D are highly consistent. The remainder of this article is organized as follows. Section II briefly introduces the theoretical framework. Detailed numerical results and discussion are given in Section III . Finally, a summary is provided in Section IV . II THEORETICAL FRAMEWORK II.1 TPA framework The half-life T_{1/2} for \alpha decay could