Pith. sign in

REVIEW 3 major objections 5 minor 1 cited by

Validation and extrapolation of atomic mass with physics-informed fully connected neural network

T0 review · 3 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read A physics-informed fully connected neural network reproduces known atomic masses to 0.122 MeV and predicts newly measured masses to 0.191 MeV.

desk verdict The paper's headline RMSDs are likely optimistically selected because early stopping and architecture choices are tuned on the same test set that is later scored, so the 0.1 MeV claim is not yet established. read the letter →

arxiv 2501.01352 v2 pith:5AJVDMDC submitted 2025-01-02 nucl-th nucl-ex

Yiming Huang , Jinhui Chen , Jiangyong Jia , Lu-Meng Liu , Yu-Gang Ma , Chunjian Zhang This is my paper · ORCID
classification nucl-thnucl-ex
keywords atomicmasspredictionfullyconnectedneuralnetworkmacroscopic-microscopicmodelresiduallearningnuclearbindingenergymagicnumbersextrapolationAME2020
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper tries to establish that a fully connected neural network, given macroscopic-microscopic features of the nucleus, can reproduce the known atomic mass surface to a test-set root-mean-square deviation of 0.122 MeV and can predict masses newly added between AME2016 and AME2020 to 0.191 MeV. The central design move is residual learning: instead of predicting the total binding energy directly, the network predicts the difference between the experimental binding energy and a liquid-drop-model baseline, while also outputting the baseline itself as an auxiliary label. If true, this would mean machine learning can act as a fast, accurate complement to mass models in regions where experimental data are sparse.

What carries the argument

The load-bearing mechanism is multi-output residual learning built on a macroscopic-microscopic baseline. The network takes liquid-drop and shell-model-inspired features, including valence nucleon numbers relative to magic numbers, and predicts three outputs: the experimental binding energy $E_{\rm exp}$, the liquid-drop model binding energy $E_{\rm LD}$, and their difference $E_{\rm exp}-E_{\rm LD}$. Training on the residual forces the network to learn the microscopic corrections that the baseline misses, which is what produces the sub-0.2 MeV accuracy.

What would settle it

Retrain Method III on the same AME2016 data with a separate validation split used only for early stopping, then evaluate once on the untouched AME2020 additions. If the resulting extrapolation RMSD is materially larger than 0.191 MeV, or if the Case A test RMSD with validation-based stopping is materially larger than 0.122 MeV, the central accuracy claim does not hold.

Watch

Extended reading notes

Core claim

The paper's central claim is that Method III, a two-hidden-layer FCNN with WS4-derived input features and three output labels, achieves a test RMSD of 0.122 MeV on AME2020 and an extrapolation RMSD of 0.191 MeV when trained on AME2016 and tested on nuclei newly added in AME2020. It further claims that sensitivity and ablation analyses show the inputs act in a physically interpretable way, and that predicted neutron and proton separation energies reproduce pairing oscillations and cusps at magic numbers, including proposed superheavy magic numbers. The authors present this as evidence that a physics-informed neural network can validate existing mass data and extend modestly beyond the measured region.

Load-bearing premise

The claimed accuracy assumes the test set was not used to tune the stopping criterion; because the early-stopping loss threshold was chosen by watching test-set error, the reported RMSD values may be optimistic.

Editorial extensions

If this is right

  • A model with test RMSD near 0.1 MeV can serve as a fast interpolator of the known mass surface, giving mass estimates for nuclei between measured points.
  • The extrapolation result suggests the same architecture can flag plausible masses for newly measured isotopes before an updated Atomic Mass Evaluation appears.
  • Pairing oscillations and magic-number kinks in predicted separation energies indicate the network can be used to search for shell closures in unexplored regions.
  • Because the method only needs a theoretical baseline and measured masses, it can be reapplied with any improved mass model as the baseline.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The reported accuracy is selected by an early-stopping rule that tracks test-set performance; an independent validation-based retraining could yield larger errors than the headline numbers.
  • The extrapolation test covers only the small set of nuclei newly added between AME2016 and AME2020, so the 0.191 MeV figure should not be read as evidence for reliable predictions deep into unknown regions like the neutron drip line.
  • A natural transfer test would be to apply the same residual-learning setup to other nuclear observables, such as charge radii or beta-decay half-lives, using the best available theoretical baseline for each.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The paper compares three fully connected neural network (FCNN) methods for atomic mass prediction. Method I is a direct regression of experimental binding energies from macroscopic and shell-model input features. Method II uses a macroscopic-microscopic decomposition with a liquid-drop baseline and multi-output training on E_exp, E_LD, and E_exp - E_LD. Method III refines Method II by using WS4-derived Coulomb and deformation features. The authors report a test-set RMSD of 0.122 MeV on AME2020 for Method III (Case A) and an extrapolation RMSD of 0.191 MeV when training on AME2016 and testing on nuclei newly added in AME2020 (Case B). They also extend the training set with WS4 theoretical masses (Case C) and use predicted separation energies to claim reproduction of pairing and magic-number effects.

Significance. If the reported accuracy holds under an unbiased evaluation, the paper would provide a useful demonstration that a physics-informed FCNN with residual multi-output learning can represent the known mass surface at the ~0.1 MeV level and extrapolate moderately beyond the training region. The design choices are broadly sensible: macroscopic-microscopic features, residual learning, auxiliary outputs, and careful preprocessing in which scaling parameters are derived from the training set only are all strengths. The main limitation is that the reported test errors are selected using test-set information, so the headline numbers are optimistic and not yet reliable as generalization estimates. The magic-number 'reproduction' claim also needs to be reframed because the magic numbers are encoded directly in the input features. With a corrected validation protocol and clarified ablation, the work would be a solid contribution.

major comments (3)
  1. [Sec. II.D and III.A] The architecture-selection rule in Sec. II.D ('if the training set performance is similar, we choose the one that performs better on the test set') and the early-stopping thresholds in Sec. III.A (e.g., 1.28e-4 for Method II and 5.97e-5 for Method III in Case A, and 1.68e-4 and 3.80e-4 in Case B) are chosen from curves of test-set RMSD versus loss shown in Fig. 4. Consequently the reported headline test RMSDs of 0.122 MeV (Case A) and 0.191 MeV (Case B) are minima over stopping criteria and architectures selected using the same nuclei on which they are then scored. They are therefore optimistically biased estimates of generalization. To support the accuracy claims, early stopping and architecture selection must be performed on a separate validation set, or via nested cross-validation, with the test set used exactly once. The current single split (random_state = 42) and the absence of released code make the size of this bias unquantifiable.
  2. [Sec. III.A (ablation paragraph)] The ablation paragraph is internally inconsistent. It reports 'For Method I, the training and test RMSD are 0.756 MeV and 0.852 MeV' and corresponding numbers for Methods II and III, but Table III Case A gives baseline values of 0.135/0.204, 0.087/0.143, and 0.052/0.122 MeV for training/test for Methods I, II, and III, respectively. The paragraph does not state which feature is removed in each reported set of numbers, and the final sentence reports Method III training/test RMSD of 0.069/0.154 MeV after removing def, which cannot be reconciled with either the 0.052/0.122 MeV baseline or the preceding 0.172/0.196 MeV numbers. Since the ablation analysis is the evidence for the claim that the δnp and def inputs are indispensable, the paragraph must be rewritten to identify each ablation and to verify the numbers.
  3. [Eq. (5) and Sec. III.C] The claimed reproduction of magic-number effects is substantially inherited from the input encoding. The valence features Vp and Vn in Eq. (5) are constructed using the prescribed magic-number list Z = 20, 28, 50, 82, 114, 120 and N = 20, 28, 50, 82, 126, 184, 198. Any regression model using these features has kinks at exactly those nucleon numbers, so the discontinuities in Sn, Sp, ΔSn, and ΔSp shown in Fig. 7 do not provide independent evidence for the magic numbers, including the newly proposed ones. The statement in Sec. IV that the three methods 'validated new magic number' is therefore overstated. To support such a claim, the authors would need to show that the network still develops the same kinks without Vp/Vn inputs, or that an alternative magic-number prescription fits notably worse; otherwise the claim should be softened to state that the model is consistent with the input magic numbers.
minor comments (5)
  1. [Sec. II.B, Eq. (1)] The symbol Enn in Eq. (1) is not defined; it should be written as E_nn or E_net to denote the neural-network prediction.
  2. [Sec. II.C.2] The text says 'Theoretical binding energies, which are highly accurate, are excluded as output labels to avoid redundancy,' but Table I lists ELD as an output for both Methods II and III. This contradiction should be resolved, for example by clarifying that only the WS4 total binding energy is meant.
  3. [Sec. III.C] The text says 'In Case A, we trained and optimized two models,' but Case A compares three methods; this sentence should be corrected to avoid implying that Method I was not part of the comparison.
  4. [Sec. III.C] The FCNN used for the separation-energy predictions in Fig. 7 is described as having 60 neurons in the first hidden layer and 30 in the second, whereas Case A models use 80/40 and Case B models use 120/60; the text should specify which trained model generated Fig. 7 and why a different architecture was used.
  5. [Sec. IV] There is a typo in 'Baysesian emulators'; it should be 'Bayesian emulators.'

Circularity Check

3 steps flagged · score 6.0 of 10

Reported test RMSDs are selected via test-set-based early stopping; magic-number reproduction is built into input features and WS4 training data.

  1. self definitional [Sec. II.C, Eq. (5); Sec. III.C/Fig. 7; Sec. IV]
    "Vp = Z − Zcore, Vn = N − Ncore, (5) where Zcore and Ncore denote the nearest magic numbers smaller than the given Z and N. In this study, the magic number adopted is (Z = 20, 28, 50, 82, 114, 120 and N = 20, 28, 50, 82, 126, 184, 198), based on the reference [68–70]."

    The model's 'reproduction' of magic-number effects is a direct consequence of the input encoding: Vp and Vn change sharply at exactly the prescribed magic numbers, so the FCNN output will show kinks at those Z/N values. The paper then presents these kinks as independent validation ('Abrupt changes near the magic number indicate the enhanced stability of magic nuclei') and in the Summary claims the methods 'validated new magic number.' Because the new magic numbers (114, 120, 184, 198) were inserted as input features by construction rather than discovered from the data, the validation claim is self-definitional.

  2. fitted input called prediction [Sec. II.D and Sec. III.A, Fig. 4; Tab. III]
    "Our model's network structure selection criterion is that if the training set performance is similar, we choose the one that performs better on the test set. ... For Method II, training was halted when the loss reached a threshold of 1.28 × 10−4 after nearly 30000 epochs, as further reduction did not consistently improve the performance of the test set."

    The headline accuracies (Method III test RMSD 0.122 MeV in Case A and 0.191 MeV in Case B) are produced after the network architecture and early-stopping thresholds are chosen by watching the test-set RMSD, as shown in Fig. 4. The same test nuclei are used both to select the stopping point and to score the claim, so the quoted RMSDs are optimistic minima over hyperparameters rather than unbiased estimates of generalization. This is a fitted input (the stopping threshold) being reported as a prediction, making the central numerical claim statistically forced.

1 more flagged steps
  1. other [Sec. III.C; Sec. IV]
    "Using AME2020 data, we incorporated theoretical values from the WS4 model as auxiliary inputs to assist in the extrapolation process in the Case C study. ... By incorporating predictions from the WS4 model, three methods validated new magic number and provided insights into superheavy nuclei."

    In Case C, WS4 theoretical masses are added directly to the training data, and the same WS4 predictions are then used as the benchmark for agreement and as the basis for claiming that new magic numbers are 'validated.' The extrapolation agreement with WS4 is therefore a check against the model's own training target, not an independent confirmation. Combined with the fact that the new magic numbers are already encoded in the Vp/Vn features, the physical-validation claim reduces to re-stating the priors that were put into the network.

full rationale

The paper's core numerical claims (Method III test RMSD 0.122 MeV in Case A and extrapolation RMSD 0.191 MeV in Case B) are not fully independent because the architecture and early-stopping thresholds are selected using the test set itself (Sec. II.D and Sec. III.A, Fig. 4). This makes the reported test errors selected minima rather than unbiased held-out errors. Separately, the paper's physical-interpretation claims are circular: magic-number effects are injected through the Vp/Vn input features with a prescribed magic-number list, and the WS4 agreement in Case C is produced by training on WS4 theoretical values and then comparing with WS4. These are genuine reductions of a claimed 'validation' to the model's own inputs. There is no self-citation chain or imported uniqueness theorem involved; the derivation of mass predictions from AME data is otherwise a standard regression exercise. The circularity is partial but load-bearing for the two headline achievements: the reported accuracy numbers and the claimed independent reproduction of magic numbers and WS4 extrapolations.

Assumptions & free parameters 6 free parameters · 5 assumptions · 0 invented entities

The model imports its physical content from the liquid-drop model and the WS4 mass formula rather than deriving it. The main free parameters are network and stopping choices tuned on the test set. No new physical entities are introduced.

free parameters (6)
  • Hidden layer sizes = 80/40 (Case A), 120/60 (Case B), 60/30 (Case C)
    Architecture sizes were selected by preliminary experiments balancing training and test performance; no separate validation set was used.
  • Early stopping loss thresholds = 1.28e-4, 5.97e-5, 1.68e-4, 3.80e-4
    Chosen from test-set error curves in Fig. 4, which introduces selection bias into the reported RMSD values.
  • Learning rate schedules = 3e-3 down to 3e-6 depending on loss
    Hand-adjusted based on loss behavior (Tab. II), not derived from data.
  • L2 regularization coefficient = 5e-8 (Case A Method I), 5e-7 (Case B Method I)
    Chosen manually to ensure convergence; not used for Methods II and III.
  • Magic number list = Z=20,28,50,82,114,120; N=20,28,50,82,126,184,198
    Assumed shell closures imported from refs; used to construct Vp and Vn, so it controls where shell effects appear in predictions.
  • Data split random state = 42
    Single fixed split; no multiple seeds or cross-validation are reported.
assumptions (5)
  • domain assumption Liquid-drop mass formula Eq. (3)-(4) with coefficients from Ref. [66] describes the macroscopic binding energy.
    Used to construct the ELD label and input features in Methods II and III; if inaccurate, the residual target changes.
  • domain assumption The shell closures listed in Section II.C are valid magic numbers across the nuclear chart.
    Defines Vp and Vn; underlies the claimed magic-number reproduction.
  • domain assumption WS4 model parameters and deformation values from Ref. [24] are reliable inputs for the neural network.
    Method III uses WS4 Coulomb form, deformation def, and ELD parameters; extrapolated-region agreement with WS4 is therefore partly by construction.
  • domain assumption AME2020 and AME2016 experimental masses with errors below 100 keV are accurate enough to serve as training labels.
    Labels are taken as ground truth; random measurement errors are not propagated into the reported RMSD.
  • standard math Neural network function classes with two hidden layers and GELU activation can approximate the mass surface.
    Relies on universal approximation; standard ML assumption.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Validation and extrapolation of atomic mass with physics-informed fully connected neural network." pith.science (2026). https://pith.science/paper/5AJVDMDC

@misc{pith2026250101352,
  author       = {Pith},
  title        = {Pith review of: Validation and extrapolation of atomic mass with physics-informed fully connected neural network},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5AJVDMDC}},
  note         = {Machine review of arXiv:2501.01352}
}
read the original abstract

Machine learning offers a powerful framework for validating and predicting atomic mass. We compare three improved neural network methods for representation and extrapolation for atomic mass prediction. The powerful method, adopting a macroscopic-microscopic approach and treating complex nuclear effects as output labels, achieves superior accuracy in AME2020, yielding a much lower root-mean-square deviation of 0.122 MeV in the test set, significantly lower than alternative methods. It also exhibits a better extrapolation performance when predicting AME2020 from AME2016, with a root-mean-square deviation of 0.191 MeV. We further conduct sensitivity analyses against the model inputs to verify interpretable alignment beyond statistical metrics. Incorporating theoretical predictions of magic numbers and masses, our fully connected neural networks reproduce key nuclear phenomena including nucleon pairing correlation and magic number effects. The extrapolation capability of the framework is discussed and the accuracy of predicting new mass measurements for isotope chains has also been tested.

Figures

Figures reproduced from arXiv: 2501.01352 by the authors.

Figure 1
Figure 1. FIG. 1. (a) The mass data were selected from the AME2020 and AME2016 datasets, with [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. FIG. 2. (a) The single-layer perceptron consists of an in [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. FIG. 3. Comparison of normalized sensitivity analysis in dif [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: FIG. 4. Variation of the RMSD of the test set with the values [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]
Figure 5
Figure 5. Figure 5: FIG. 5. Prediction errors of three Methods on the AME2020 [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 6
Figure 6. Figure 6: FIG. 6. Comparison of the residual in three Methods in newly [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]
Figure 7
Figure 7. Figure 7: FIG. 7. Predicted neutron ( [PITH_FULL_IMAGE:figures/full_fig_p009_7.png]
Figure 8
Figure 8. Figure 8: FIG. 8. Comparison of extrapolation in different mass models for (a) [PITH_FULL_IMAGE:figures/full_fig_p010_8.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Heavy Quarkonium Spectrum and Decay Constants from a Neural-Network-Based Holographic Model

    hep-ph 2026-01 conditional novelty 5.0 of 10

    A neural-network-parametrized dilaton field reproduces the masses and leptonic decay constants of charmonium and bottomonium with 1.26% and 3.32% RMS errors, but only because those values were used as training data.

Reference graph

Works this paper leans on

94 extracted references · 73 canonical work pages · cited by 1 Pith paper

  1. [1]

    Input and output features in Method I, II, and III

    Method I In this method, we consider versions of the liquid droplet model where the pairing term takes the form 4 TABLE I. Input and output features in Method I, II, and III. Methods Input Features Output Features I Z, A, A2/3, EC /ac, I2A, |I|, A−1/3, Vp, Vn, δnp, A−1/2 Eexp II Z, A, A2/3, EC /ac, I2A, |I|, A−1/3, Vp, Vn, δnp, g1A1/3, g2A−1/3 ELD, Eexp, ...

  2. [2]

    Method II To enhance prediction accuracy, we developed an alter- native approach, Method II, which incorporates revised input features, output features, and network architecture for comparison with Method I. The total binding energy of a nucleus, which includes deformation and Strutinsky shell corrections, E′(A, Z, β), is expressed as E(A, Z, β) = ELD(A, ...

  3. [3]

    black box

    Method III In this method, we adopt a strategy similar to Method II but with a few minor adjustments, as listed in Tab. I (Method III). Specifically, the Coulomb term differs from that in Eq. (4) and takes the following form, as used in the WS4 model [24]. EC = ac Z 2 A1/3 1 − 0.76Z −2/3 . (9) Next, the parameter δnp is encoded as follows δnp =  ...

    1900

  4. [4]

    Recent trends in the determination of nuclear masses,

    D. Lunney, J. M. Pearson, and C. Thibault, “Recent trends in the determination of nuclear masses,” Rev. Mod. Phys. 75, 1021–1082 (2003)

    2003

  5. [5]

    Impact of nuclear mass uncertainties on the r-process,

    Dirk Martin, Almudena Arcones, Witold Nazarewicz, and Erik Olsen, “Impact of nuclear mass uncertainties on the r-process,” Phys. Rev. Lett. 116, 121101 (2016)

    2016

  6. [6]

    Modern Theory of Nuclear Forces,

    Evgeny Epelbaum, Hans-Werner Hammer, and Ulf- G. Meissner, “Modern Theory of Nuclear Forces,” Rev. Mod. Phys. 81, 1773–1825 (2009)

    2009

  7. [7]

    The Equation of state of nucleon matter and neu- tron star structure,

    A. Akmal, V. R. Pandharipande, and D. G. Raven- hall, “The Equation of state of nucleon matter and neu- tron star structure,” Phys. Rev. C 58, 1804–1828 (1998), arXiv:nucl-th/9804027

    arXiv 1998

  8. [8]

    Equation of state of stellar nuclear matter and the effective nucleon mass,

    M. Onsi and J. M. Pearson, “Equation of state of stellar nuclear matter and the effective nucleon mass,” Phys. Rev. C 65, 047302 (2002)

    2002

Show all 94 references
  1. [9]

    Nuclear masses and deformations,

    William D. Myers and Wladyslaw J. Swiatecki, “Nuclear masses and deformations,” Nucl. Phys. 81, 1–60 (1966)

    1966

  2. [10]

    Synthesis of the elements in stars,

    Margaret E. Burbidge, G. R. Burbidge, William A. Fowler, and F. Hoyle, “Synthesis of the elements in stars,” Rev. Mod. Phys. 29, 547–650 (1957)

    1957

  3. [11]

    The impact of individual nuclear prop- erties on r-process nucleosynthesis,

    M. R. Mumpower, R. Surman, G. C. McLaughlin, and A. Aprahamian, “The impact of individual nuclear prop- erties on r-process nucleosynthesis,” Prog. Part. Nucl. Phys. 86, 86–126 (2016), [Erratum: Prog. Part. Nucl. Phys. 87, 116–116 (2016)]

    2016

  4. [12]

    rp-process nucleosynthesis at extreme temperature and density conditions,

    H. Schatz et al., “rp-process nucleosynthesis at extreme temperature and density conditions,” Phys. Rep. 294, 167–263 (1998)

    1998

  5. [13]

    Location of the Neutron Dripline at Fluorine and Neon,

    D. S. Ahn et al., “Location of the Neutron Dripline at Fluorine and Neon,” Phys. Rev. Lett. 123, 212501 (2019)

    2019

  6. [14]

    The limits of the nuclear landscape,

    Jochen Erler, Noah Birge, Markus Kortelainen, Witold Nazarewicz, Erik Olsen, Alexander M. Perhac, and Mario Stoitsov, “The limits of the nuclear landscape,” Nature 486, 509–512 (2012)

    2012

  7. [15]

    The History of nuclidic masses and of their evaluation,

    Georges Audi, “The History of nuclidic masses and of their evaluation,” Int. J. Mass Spectr. Ion Process. 251, 85–94 (2006)

    2006

  8. [16]

    The Ame 2003 atomic mass evaluation,

    A. H. Wapstra, G. Audi, and C. Thibault, “The Ame 2003 atomic mass evaluation,” Nucl. Phys. A 729, 129– 336 (2003)

    2003

  9. [17]

    The Ame2003 atomic mass evaluation (II). Tables, graphs and references,

    G. Audi, A. H. Wapstra, and C. Thibault, “The Ame2003 atomic mass evaluation (II). Tables, graphs and references,” Nucl. Phys. A 729, 337–676 (2002)

    2002

  10. [18]

    The Ame2012 atomic mass evaluation,

    Wapstra A.H. et al. Audi G., Wang M., “The Ame2012 atomic mass evaluation,” Chin. Phys. C 36, 1287 (2012)

    2012

  11. [19]

    The Ame2012 atomic mass evaluation,

    M. Wang, G. Audi, A. H. Wapstra, F. G. Kondev, M. MacCormick, X. Xu, and B. Pfeiffer, “The Ame2012 atomic mass evaluation,” Chin. Phys. C 36, 1603–2014 (2012)

    2012

  12. [20]

    The AME2016 atomic mass evaluation (I). Evaluation of in- put data; and adjustment procedures,

    W.J. Huang, G. Audi, M. Wang, and et al., “The AME2016 atomic mass evaluation (I). Evaluation of in- put data; and adjustment procedures,” Chin. Phys. C 41, 030002 (2017)

    2017

  13. [21]

    Nuclear Physics A. Sta- tionary States of Nuclei,

    H. A. Bethe and R. F. Bacher, “Nuclear Physics A. Sta- tionary States of Nuclei,” Rev. Mod. Phys. 8, 82–229 (1936)

    1936

  14. [22]

    Zur Theorie der Kernmassen,

    C. F. V. Weizsacker, “Zur Theorie der Kernmassen,” Z. Phys. 96, 431–458 (1935)

    1935

  15. [23]

    Nuclear mass predictions with machine learning reaching the accuracy required by r-process studies,

    Z. M. Niu and H. Z. Liang, “Nuclear mass predictions with machine learning reaching the accuracy required by r-process studies,” Phys. Rev. C 106, L021303 (2022)

    2022

  16. [24]

    Nuclear ground-state masses and deformations: FRDM(2012),

    P. M¨ oller, A. J. Sierk, T. Ichikawa, and H. Sagawa, “Nuclear ground-state masses and deformations: FRDM(2012),” Atom. Data Nucl. Data Tabl. 109-110, 1–204 (2016)

    2012

  17. [25]

    Nuclear mass formula via an approximation to the Hartree—Fock method,

    Y. Aboussir, J. M. Pearson, A. K. Dutta, and F. Ton- deur, “Nuclear mass formula via an approximation to the Hartree—Fock method,” Atom. Data Nucl. Data Tabl. 61, 127–176 (1995)

    1995

  18. [26]

    Further ex- plorations of Skyrme-Hartree-Fock-Bogoliubov mass for- mulas. 13. The 2012 atomic mass evaluation and the sym- 12 metry coefficient,

    S. Goriely, N. Chamel, and J. M. Pearson, “Further ex- plorations of Skyrme-Hartree-Fock-Bogoliubov mass for- mulas. 13. The 2012 atomic mass evaluation and the sym- 12 metry coefficient,” Phys. Rev. C 88, 024308 (2013)

    2013

  19. [27]

    Sur- face diffuseness correction in global mass formula,

    Ning Wang, Min Liu, Xizhen Wu, and Jie Meng, “Sur- face diffuseness correction in global mass formula,” Phys. Lett. B 734, 215–219 (2014)

    2014

  20. [28]

    Relativistic mean field in finite nuclei,

    P. Ring, “Relativistic mean field in finite nuclei,” Prog. Part. Nucl. Phys. 37, 193–263 (1996)

    1996

  21. [29]

    Relativistic Hartree Bogoliubov theory: static and dynamic aspects of exotic nuclear structure,

    D. Vretenar, A. V. Afanasjev, G. A. Lalazissis, and P. Ring, “Relativistic Hartree Bogoliubov theory: static and dynamic aspects of exotic nuclear structure,” Phys. Rep. 409, 101–259 (2005)

    2005

  22. [30]

    Multidimensionally constrained covari- ant density functional theories—nuclear shapes and po- tential energy surfaces,

    Shan-Gui Zhou, “Multidimensionally constrained covari- ant density functional theories—nuclear shapes and po- tential energy surfaces,” Phys. Scripta91, 063008 (2016)

    2016

  23. [31]

    The limits of the nuclear landscape ex- plored by the relativistic continuum Hartree–Bogoliubov theory,

    X. W. Xia et al., “The limits of the nuclear landscape ex- plored by the relativistic continuum Hartree–Bogoliubov theory,” Atom. Data Nucl. Data Tabl. 121-122, 1–215 (2018)

    2018

  24. [32]

    New parametrization for the nuclear covariant energy density functional with point-coupling interaction,

    P. W. Zhao, Z. P. Li, J. M. Yao, and J. Meng, “New parametrization for the nuclear covariant energy density functional with point-coupling interaction,” Phys. Rev. C 82, 054319 (2010)

    2010

  25. [33]

    Nuclear mass table in deformed relativistic Hartree–Bogoliubov theory in continuum, I: Even–even nuclei,

    Kaiyuan Zhang et al. (DRHBc Mass Table), “Nuclear mass table in deformed relativistic Hartree–Bogoliubov theory in continuum, I: Even–even nuclei,” Atom. Data Nucl. Data Tabl. 144, 101488 (2022)

    2022

  26. [34]

    Nuclear mass table in deformed relativistic Hartree–Bogoliubov theory in continuum, II: Even-Z nuclei,

    Peng Guo et al. (DRHBc Mass Table), “Nuclear mass table in deformed relativistic Hartree–Bogoliubov theory in continuum, II: Even-Z nuclei,” Atom. Data Nucl. Data Tabl. 158, 101661 (2024)

    2024

  27. [35]

    Refining mass formulas for astrophysical applications: a Bayesian neu- ral network approach,

    Raditya Utama and Jorge Piekarewicz, “Refining mass formulas for astrophysical applications: a Bayesian neu- ral network approach,” Phys. Rev. C 96, 044308 (2017)

    2017

  28. [36]

    Global prediction of nuclear charge density dis- tributions using a deep neural network,

    Tian Shuai Shang, Hui Hui Xie, Jian Li, and Haozhao Liang, “Global prediction of nuclear charge density dis- tributions using a deep neural network,” Phys. Rev. C 110, 014308 (2024)

    2024

  29. [37]

    Nuclear mass predictions based on convolutional neural network,

    Yanhua Lu, Tianshuai Shang, Pengxiang Du, Jian Li, Haozhao Liang, and Zhongming Niu, “Nuclear mass predictions based on convolutional neural network,” arXiv:2404.14948 [nucl-th]

  30. [38]

    Validating neural- network refinements of nuclear mass models,

    R. Utama and J. Piekarewicz, “Validating neural- network refinements of nuclear mass models,” Phys. Rev. C 97, 014306 (2018)

    2018

  31. [39]

    Deep learning approach to nuclear masses and α-decay half-lives,

    Chen-Qi Li, Chao-Nan Tong, Hong-Jing Du, and Long- Gang Pang, “Deep learning approach to nuclear masses and α-decay half-lives,” Phys. Rev. C 105, 064306 (2022)

    2022

  32. [40]

    Nuclear binding energies in artificial neural networks,

    Lin-Xing Zeng, Yu-Ying Yin, Xiao-Xu Dong, and Li- Sheng Geng, “Nuclear binding energies in artificial neural networks,” Phys. Rev. C 109, 034318 (2024)

    2024

  33. [41]

    Predictions of nuclear charge radii and physical interpretations based on the naive Bayesian probability classifier,

    Yunfei Ma, Chen Su, Jian Liu, Zhongzhou Ren, Chang Xu, and Yonghao Gao, “Predictions of nuclear charge radii and physical interpretations based on the naive Bayesian probability classifier,” Phys. Rev. C 101, 014304 (2020)

    2020

  34. [42]

    Novel Bayesian neural network based approach for nuclear charge radii,

    Xiao-Xu Dong, Rong An, Jun-Xu Lu, and Li-Sheng Geng, “Novel Bayesian neural network based approach for nuclear charge radii,” Phys. Rev. C 105, 014308 (2022)

    2022

  35. [43]

    Predictions of nuclearβ -decay half-lives with machine learning and their impact on r -process nucle- osynthesis,

    Z. M. Niu, H. Z. Liang, B. H. Sun, W. H. Long, and Y. F. Niu, “Predictions of nuclearβ -decay half-lives with machine learning and their impact on r -process nucle- osynthesis,” Phys. Rev. C 99, 064307 (2019)

    2019

  36. [44]

    Bayesian optimization approach to model-based description of α decay,

    Zisheng Jin, Mingshuai Yan, Hao Zhou, An Cheng, Zhongzhou Ren, and Jian Liu, “Bayesian optimization approach to model-based description of α decay,” Phys. Rev. C 108, 014326 (2023)

    2023

  37. [45]

    Simple deep-learning approach for α-decay half-life studies,

    Na-Na Ma, Tian-Liang Zhao, Wen-Xia Wang, and Hong- Fei Zhang, “Simple deep-learning approach for α-decay half-life studies,” Phys. Rev. C 107, 014310 (2023)

    2023

  38. [46]

    Comprehensive estimation of nuclide production cross sections using a phenomenological approach,

    Hiroki Iwamoto, Shin-ichiro Meigo, and Kenta Sugihara, “Comprehensive estimation of nuclide production cross sections using a phenomenological approach,” Phys. Rev. C 109, 054610 (2024)

    2024

  39. [47]

    Unmasking Cor- relations in Nuclear Cross Sections with Graph Neural Networks,

    Sinjini Mitra, Hongjun Choi, Shusen Liu, Ruben Glatt, Kyle Wendt, and Nicolas Schunck, “Unmasking Cor- relations in Nuclear Cross Sections with Graph Neural Networks,” arXiv:2404.02332 [nucl-th]

  40. [48]

    Transfer learning and neural networks in predicting quadrupole deformation,

    Yuan Lin, Jia-Xing Li, and Hong-Fei Zhang, “Transfer learning and neural networks in predicting quadrupole deformation,” Chin. Phys. C 48, 064106 (2024)

    2024

  41. [49]

    Mapping low-lying states and B(E2;01+→21+) in even- even nuclei with machine learning,

    B. F. Lv, Z. L. Li, Y. J. Wang, and C. M. Petrache, “Mapping low-lying states and B(E2;01+→21+) in even- even nuclei with machine learning,” Phys. Lett. B 857, 139013 (2024)

    2024

  42. [50]

    A neural network approach for orienting heavy-ion collision events,

    Zu-Xing Yang, Xiao-Hua Fan, Zhi-Pan Li, and Shunji Nishimura, “A neural network approach for orienting heavy-ion collision events,” Phys. Lett. B 848, 138359 (2024)

    2024

  43. [51]

    Phase Transition Study Meets Machine Learn- ing,

    Yu-Gang Ma, Long-Gang Pang, Rui Wang, and Kai Zhou, “Phase Transition Study Meets Machine Learn- ing,” Chin. Phys. Lett. 40, 122101 (2023)

    2023

  44. [52]

    High-energy nuclear physics meets machine learning,

    Wan-Bing He, Yu-Gang Ma, Long-Gang Pang, Hui-Chao Song, and Kai Zhou, “High-energy nuclear physics meets machine learning,” Nucl. Sci. Tech. 34, 88 (2023)

    2023

  45. [53]

    Exploring QCD matter in extreme conditions with Machine Learning,

    Kai Zhou, Lingxiao Wang, Long-Gang Pang, and Shuzhe Shi, “Exploring QCD matter in extreme conditions with Machine Learning,” Prog. Part. Nucl. Phys. 135, 104084 (2024)

    2024

  46. [54]

    Machine learning study to identify collective flow in small and large colliding systems,

    Shuang Guo, Han-Sheng Wang, Kai Zhou, and Guo- Liang Ma, “Machine learning study to identify collective flow in small and large colliding systems,” Phys. Rev. C 110, 024910 (2024)

    2024

  47. [55]

    Nuclear liquid-gas phase transition with machine learning,

    Rui Wang, Yu-Gang Ma, R. Wada, Lie-Wen Chen, Wan- Bing He, Huan-Ling Liu, and Kai-Jia Sun, “Nuclear liquid-gas phase transition with machine learning,” Phys. Rev. Res. 2, 043202 (2020)

    2020

  48. [56]

    Determining the temperature in heavy-ion collisions with multiplicity distribution,

    Yi-Dan Song, Rui Wang, Yu-Gang Ma, Xian-Gai Deng, and Huan-Ling Liu, “Determining the temperature in heavy-ion collisions with multiplicity distribution,” Phys. Lett. B 814, 136084 (2021), arXiv:2101.10613 [nucl-th]

    2021 arXiv

  49. [57]

    Properties of the QCD matter: re- view of selected results from the relativistic heavy ion collider beam energy scan (RHIC BES) program,

    Jinhui Chen et al., “Properties of the QCD matter: re- view of selected results from the relativistic heavy ion collider beam energy scan (RHIC BES) program,” Nucl. Sci. Tech. 35, 214 (2024)

    2024

  50. [58]

    Properties of QCD matter: a review of selected results from ALICE experiment,

    Qi-Ye Shou et al., “Properties of QCD matter: a review of selected results from ALICE experiment,” Nucl. Sci. Tech. 35, 219 (2024)

    2024

  51. [59]

    Studies on several prob- lems in nuclear physics by using machine learning,

    Zepeng Gao and Qinggeng Li, “Studies on several prob- lems in nuclear physics by using machine learning,” Nucl. Tech. 46, 080009 (2023)

    2023

  52. [60]

    Machine learning in nuclear physics at low and intermediate energies,

    Wanbing He, Qingfeng Li, Yugang Ma, Zhongming Niu, Junchen Pei, and Yingxun Zhang, “Machine learning in nuclear physics at low and intermediate energies,” Sci. China Phys. Mech. Astron. 66, 282001 (2023)

    2023

  53. [61]

    Nuclear binding energy predictions using neural net- works: Application of the multilayer perceptron,

    Esra Y¨ uksel, Derya Soydaner, and H¨ useyin Bahtiyar, “Nuclear binding energy predictions using neural net- works: Application of the multilayer perceptron,” Int. 13 J. Mod. Phys. E 30, 2150017 (2021)

    2021

  54. [62]

    Correction to: Ma- chine learning the nuclear mass,

    Zepeng Gao, Yongjia Wang, Hongliang L¨ u, Qingfeng Li, Caiwan Shen, and Ling Liu, “Correction to: Ma- chine learning the nuclear mass,” Nucl. Sci. Tech. 32, 118 (2021)

    2021

  55. [63]

    Nuclear mass predictions using machine learning mod- els,

    Esra Y¨ uksel, Derya Soydaner, and H¨ useyin Bahtiyar, “Nuclear mass predictions using machine learning mod- els,” Phys. Rev. C 109, 064322 (2024)

    2024

  56. [64]

    Bayesian approach to model-based ex- trapolation of nuclear observables,

    L´ eo Neufcourt, Yuchen Cao, Witold Nazarewicz, and Frederi Viens, “Bayesian approach to model-based ex- trapolation of nuclear observables,” Phys. Rev. C 98, 034318 (2018)

    2018

  57. [65]

    On the rate of con- vergence of fully connected deep neural network regres- sion estimates,

    Michael Kohler and Sophie Langer, “On the rate of con- vergence of fully connected deep neural network regres- sion estimates,” Ann. Stat. 49, 2231 – 2249 (2021)

    2021

  58. [66]

    Multilayer feedforward networks are universal approxi- mators,

    Kurt Hornik, Maxwell Stinchcombe, and Halbert White, “Multilayer feedforward networks are universal approxi- mators,” Neural Networks 2, 359–366 (1989)

    1989

  59. [67]

    The AME 2020 atomic mass evaluation (II). Tables, graphs and references,

    Meng Wang, W. J. Huang, F. G. Kondev, G. Audi, and S. Naimi, “The AME 2020 atomic mass evaluation (II). Tables, graphs and references,” Chin. Phys. C45, 030003 (2021)

    2021

  60. [68]

    The AME2016 atomic mass evaluation (II). Tables, graphs and references,

    Meng Wang, G. Audi, F. G. Kondev, W. J. Huang, S. Naimi, and Xing Xu, “The AME2016 atomic mass evaluation (II). Tables, graphs and references,” Chin. Phys. C 41, 030003 (2017)

    2017

  61. [69]

    Mirror nuclei constraint in mass formula,

    Ning Wang, Zuoying Liang, Min Liu, and Xizhen Wu, “Mirror nuclei constraint in mass formula,” Phys. Rev. C 82, 044304 (2010)

    2010

  62. [70]

    Microscopic mass formulae,

    J. Duflo and A. P. Zuker, “Microscopic mass formulae,” Phys. Rev. C 52, R23 (1995), arXiv:nucl-th/9505011

    1995 arXiv

  63. [71]

    Nuclear magic numbers: New features far from stability,

    O. Sorlin and M. G. Porquet, “Nuclear magic numbers: New features far from stability,” Prog. Part. Nucl. Phys. 61, 602–673 (2008), arXiv:0805.2561 [nucl-ex]

    2008 arXiv

  64. [72]

    Self-consistent mean-field models for nuclear structure,

    Michael Bender, Paul-Henri Heenen, and Paul-Gerhard Reinhard, “Self-consistent mean-field models for nuclear structure,” Rev. Mod. Phys. 75, 121–180 (2003)

    2003

  65. [73]

    Relativistic Continuum Hartree Bogoli- ubov theory for ground state properties of exotic nuclei,

    J. Meng, H. Toki, S. G. Zhou, S. Q. Zhang, W. H. Long, and L. S. Geng, “Relativistic Continuum Hartree Bogoli- ubov theory for ground state properties of exotic nuclei,” Prog. Part. Nucl. Phys. 57, 470–563 (2006), arXiv:nucl- th/0508020

    2006

  66. [74]

    261 (Springer, 1996)

    Walter Greiner, Joachim A Maruhn, et al., Nuclear mod- els, Vol. 261 (Springer, 1996)

    1996

  67. [75]

    Learning from noisy labels with deep neural networks: A survey,

    Hwanjun Song, Minseok Kim, Dongmin Park, Yooju Shin, and Jae-Gil Lee, “Learning from noisy labels with deep neural networks: A survey,” IEEE T. Neur. Net. Lear. 34, 8135–8153 (2022)

    2022

  68. [76]

    Residual correla- tion in graph neural network regression,

    Junteng Jia and Austion R Benson, “Residual correla- tion in graph neural network regression,” in Proceedings of the 26th ACM SIGKDD international conference on knowledge discovery & data mining(2020) pp. 588–598

    2020

  69. [77]

    Gaussian error lin- ear units (gelus),

    Dan Hendrycks and Kevin Gimpel, “Gaussian error lin- ear units (gelus),” arXiv:1606.08415 (2016)

    2016 arXiv

  70. [78]

    Gelu activation function in deep learn- ing: a comprehensive mathematical analysis and perfor- mance,

    Minhyeok Lee, “Gelu activation function in deep learn- ing: a comprehensive mathematical analysis and perfor- mance,” arXiv:2305.12073 (2023)

    2023 arXiv

  71. [79]

    Robust Estimation of a Location Pa- rameter,

    Peter J. Huber, “Robust Estimation of a Location Pa- rameter,” The Annals of Mathematical Statistics 35, 73 – 101 (1964)

    1964

  72. [80]

    Bayesian infer- ence of the crust-core transition density via the neutron- star radius and neutron-skin thickness data,

    Wenjie Xie, Ma Ziwei, and Guo Junhua, “Bayesian infer- ence of the crust-core transition density via the neutron- star radius and neutron-skin thickness data,” Nucl. Sci. Tech. 34, 91 (2023)

    2023

  73. [81]

    Bayesian inference of neutron- skin thickness and neutron-star observables based on ef- fective nuclear interactions,

    Jia Zhou and Jun Xu, “Bayesian inference of neutron- skin thickness and neutron-star observables based on ef- fective nuclear interactions,” Sci. China Phys. Mech. As- tron. 67, 282011 (2024)

    2024

  74. [82]

    Bayesian model averaging (BMA) for nuclear data evaluation,

    E. Alhassan, D. Rochman, G. Schnabel, and A. J. Kon- ing, “Bayesian model averaging (BMA) for nuclear data evaluation,” Nucl. Sci. Tech. 35, 205 (2024)

    2024

  75. [83]

    Impact parameter dependence of anisotropic flow: Bayesian reconstruction in ultracentral nucleus-nucleus collisions,

    Mubarak Alqahtani, Rajeev S. Bhalerao, Giuliano Gi- acalone, Andreas Kirchner, and Jean-Yves Ollitrault, “Impact parameter dependence of anisotropic flow: Bayesian reconstruction in ultracentral nucleus-nucleus collisions,” Phys. Rev. C 110, 064906 (2024)

    2024

  76. [84]

    Measurement of the mass dif- ference and the binding energy of the hypertriton and antihypertriton,

    J. Adam et al. (STAR), “Measurement of the mass dif- ference and the binding energy of the hypertriton and antihypertriton,” Nature Phys. 16, 409–412 (2020)

    2020

  77. [85]

    Imaging shapes of atomic nuclei in high-energy nuclear collisions,

    M. I. Abdulhamid et al. (STAR), “Imaging shapes of atomic nuclei in high-energy nuclear collisions,” Nature 635, 67–72 (2024)

    2024

  78. [86]

    Imaging the initial condition of heavy-ion collisions and nuclear structure across the nu- clide chart,

    Jiangyong Jia et al., “Imaging the initial condition of heavy-ion collisions and nuclear structure across the nu- clide chart,” Nucl. Sci. Tech. 35, 220 (2024)

    2024

  79. [87]

    Evidence of Quadrupole and Octupole Deformations in Zr 96+Zr96 and Ru96+Ru96 Collisions at Ultrarelativistic Energies,

    Chunjian Zhang and Jiangyong Jia, “Evidence of Quadrupole and Octupole Deformations in Zr 96+Zr96 and Ru96+Ru96 Collisions at Ultrarelativistic Energies,” Phys. Rev. Lett. 128, 022301 (2022)

    2022

  80. [88]

    Impact of Nuclear Deformation on Relativistic Heavy- Ion Collisions: Assessing Consistency in Nuclear Physics across Energy Scales,

    Giuliano Giacalone, Jiangyong Jia, and Chunjian Zhang, “Impact of Nuclear Deformation on Relativistic Heavy- Ion Collisions: Assessing Consistency in Nuclear Physics across Energy Scales,” Phys. Rev. Lett. 127, 242301 (2021)

    2021

  81. [89]

    Inter- pretable deep learning for nuclear deformation in heavy ion collisions,

    Long-Gang Pang, Kai Zhou, and Xin-Nian Wang, “Inter- pretable deep learning for nuclear deformation in heavy ion collisions,” arXiv:1906.06429 [nucl-th]

    1906 arXiv

  82. [90]

    Deep-neural-network approach to solving the ab initio nuclear structure prob- lem,

    Yilong Yang and Pengwei Zhao, “Deep-neural-network approach to solving the ab initio nuclear structure prob- lem,” Phys. Rev. C 107, 034320 (2023)

    2023

  83. [91]

    Beyond axial symmetry: high- energy collisions unveil the ground-state shape of 238U,

    Giacalone Giuliano, “Beyond axial symmetry: high- energy collisions unveil the ground-state shape of 238U,” Nucl. Sci. Tech. 35, 218 (2024)

    2024

  84. [92]

    Applications of deep learning to relativistic hydrodynamics,

    Hengfeng Huang, Bowen Xiao, Ziming Liu, Zeming Wu, Yadong Mu, and Huichao Song, “Applications of deep learning to relativistic hydrodynamics,” Phys. Rev. Res. 3, 023256 (2021)

    2021

  85. [93]

    An equation-of- state-meter of quantum chromodynamics transition from deep learning,

    Long-Gang Pang, Kai Zhou, Nan Su, Hannah Petersen, Horst St¨ ocker, and Xin-Nian Wang, “An equation-of- state-meter of quantum chromodynamics transition from deep learning,” Nature Commun. 9, 210 (2018)

    2018

  86. [94]

    Deep-learning quasi-particle masses from QCD equation of state,

    Fu-Peng Li, Hong-Liang L¨ u, Long-Gang Pang, and Guang-You Qin, “Deep-learning quasi-particle masses from QCD equation of state,” Phys. Lett. B 844, 138088 (2023)

    2023

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.