Machine Learning Frameworks for Unbiased Extraction of Compton Form Factors from Multidimensional Double Deeply Virtual Compton Scattering Observables”
Author
Innocent Ogoh Echi, Mohammed Abdullahi Attahiru, Suraj Aliyu
Abstract
Double Deeply Virtual Compton Scattering (DDVCS) offers an essential avenue for three-dimensional nucleon tomography by employing dual photon virtualities (Q2>0,Q'2>0) to decouple the skewness parameter ξ from the Bjorken scaling variable xB. Unlike conventional Deeply Virtual Compton Scattering, DDVCS directly probes the off-diagonal Ebert–Radyushkin–Brodsky–Lepage domain (|x|<ξ), mapping previously unconstrained quark-antiquark pair creation dynamics. However, extracting Compton Form Factors (CFFs) from multidimensional observables like Beam Spin Asymmetries (ALU) constitutes an ill-posed inverse problem complicated by model-dependent parameterization bias, non-linear Bethe–Heitler interference, and kinematic degeneracies. Here, we present a physics-informed machine learning architecture designed for unbiased CFF extraction across multidimensional phase space. The framework embeds exact quantum electrodynamic cross-section denominators, Dispersion Relations, and Next-to-Leading Order QCD constraints directly into the loss function, eliminating arbitrary functional assumptions. Validation against simulated pseudo-data demonstrates high reconstruction fidelity for ReH (R2=0.984, residual mean μ=−0.003, σ=0.087) and stable error scaling up to 10% experimental noise. Global sensitivity analysis, feature importance mapping, and local variance profiling across xB and t domains clarify key extraction bottlenecks, notably H--E correlations. This computational platform established a rigorous, model-independent strategy for processing upcoming high-luminosity data from modern hadronic facilities and the Electron-Ion Collider.
Keywords
Double Deeply Virtual Compton Scattering (DDVCS), Generalized Parton Distributions (GPDs), Physics-Informed Machine Learning
DOI : https://doi.org/10.5281/zenodo.23204399
Full Text:
References
Belitsky, A. V., & Müller, D. (2003). Exclusive electroproduction of lepton pairs as a probe of generalized parton distributions. Physical Review Letters, 90(2), Article 022001. https://doi.org/10.1103/PhysRevLett.90.022001
Belitsky, A. V., & Radyushkin, A. V. (2005). Unraveling hadron structure with generalized parton distributions. Physics Reports, 418(1–6), 1–187. https://doi.org/10.1016/j.physrep.2005.06.002
Boër, M., & Biswas, D. (2024). Double Deeply Virtual Compton Scattering at Jefferson Lab Hall A. arXiv preprint, Article arXiv:2403.02605. https://doi.org/10.48550/arXiv.2403.02605
Collins, J. C., & Freund, A. (1999). Proof of factorization for deeply virtual Compton scattering in QCD. Physical Review D, 59(7), Article 074009. https://doi.org/10.1103/PhysRevD.59.074009
Deja, K., Martínez-Fernández, V., Pire, B., Sznajder, P., & Wagner, J. (2023). Phenomenology of double deeply virtual Compton scattering in the era of new experiments. Physical Review D, 107(9), Article 094035. https://doi.org/10.1103/PhysRevD.107.094035
Diehl, M. (2003). Generalized parton distributions. Physics Reports, 388(2–4), 41–277. https://doi.org/10.1016/j.physrep.2003.08.002
Grocholski, O., Moutarde, H., Pire, B., Sznajder, P., & Wagner, J. (2020). Data-driven study of timelike Compton scattering. European Physical Journal C, 80(2), Article 171. https://doi.org/10.1140/epjc/s10052-020-7700-9
Guidal, M., Moutarde, H., & Vanderhaeghen, M. (2013). Generalized parton distributions in the valence region from deeply virtual Compton scattering. Reports on Progress in Physics, 76(6), Article 066202. https://doi.org/10.1088/0034-4885/76/6/066202
Guidal, M., & Vanderhaeghen, M. (2003). Double deeply virtual Compton scattering with CLAS at 11 GeV. Physical Review Letters, 90(1), Article 012001. https://doi.org/10.1103/PhysRevLett.90.012001
Ji, X. (1997). Gauge-invariant decomposition of nucleon spin and its spin-off. Physical Review Letters, 78(4), 610–613. https://doi.org/10.1103/PhysRevLett.78.610
Karniadakis, G. E., Kevrekidis, I. G., Lu, L., Perdikaris, P., Wang, S., & Yang, L. (2021). Physics-informed machine learning. Nature Reviews Physics, 3(6), 422–440. https://doi.org/10.1038/s42254-021-00314-5
Müller, D., & Zhang, J. (2015). Dispersion relations for deeply virtual Compton scattering and generalized parton distributions. Physical Review D, 91(7), Article 074016. https://doi.org/10.1103/PhysRevD.91.074016
Radyushkin, A. V. (1997). None forward parton distributions. Physical Review D, 56(9), 5524–5557. https://doi.org/10.1103/PhysRevD.56.5524
Zhao, S., Camsonne, A., Marchand, D., Mazouz, M., Sparveris, N., Stepanyan, S., Voutier, E., & Zhao, Z. W. (2021). Double deeply virtual Compton scattering with positron beams at SoLID. European Physical Journal A, 57(11), Article 314. https://doi.org/10.1140/epja/s10050-021-00551-3