arXiv subjects

Wanchen Li

Publications and source records attributed to Wanchen Li.

12 recordsLinked to original sources

arXiv2026

Recoil Geometry Unmasks Gluon Saturation in Forward $Z^0$ Production

Gluon saturation produces characteristic transverse-momentum broadening in nuclei, but QCD radiation largely washes out this signature. We show that fiducial recoil subtraction turns detector acceptance into a transverse-momentum projector that unmasks the broadening in forward $Z^0$ production. Subtracting the hadronic recoil measured in a chosen rapidity interval from the boson transverse momentum defines a residual momentum. At leading power, the radiative recoil in this interval cancels, while the residual momentum retains sensitivity to the small-$x$ nuclear field. Combining a CGC description of the small-$x$ target with soft-collinear effective theory (SCET) resummation for finite rapidity coverage, we find that a benchmark rapidity coverage $|\eta^{\rm lab}|<2.5$ lowers the effective hard scale from $M_Z\simeq 91.2$ GeV to about $7.5~\mathrm{GeV}$ of the Sudakov evolution. Increasing the saturation scale broadens the residual-momentum distribution and weakens recoil alignment, whereas wider coverage makes the proton--nucleus separation clearer in both observables. Detector geometry thus provides tunable control over perturbative recoil, enabling a probe of nonlinear small-$x$ QCD.

hep-ph↗

arXiv2026

Tracing Vacuum Hadronization with Conserved Currents

We show that color triality constrains the nonperturbative states that screen a Wilson-line endpoint, allowing the joint flows of net electric charge, baryon number, and strangeness to probe QCD vacuum hadronization. Whether evaluated from resolved hadrons in a jet initiated by a quark of flavor $f$ or from the corresponding charge correlators, these flows satisfy the Gell-Mann--Nishijima relation $\langle Q_{\rm flow}\rangle_f \simeq I_{3,f} +\frac{1}{2}\left( \langle S_{\rm flow}\rangle_f +\langle B_{\rm flow}\rangle_f \right)$, where $I_{3,f}$ is the third component of the initiating-quark isospin. The net jet baryon number, $\langle B_{\rm flow}\rangle_f\simeq \frac{r_{qq}}{1+r_{qq}}$, directly probes the relative probability of a diquark--antidiquark vacuum excitation, with $r_{qq}$ the diquark-to-quark production ratio. Thus, for $r_{qq}\ll1$, the baryon number carried by the jet is substantially suppressed relative to the initiating-quark value $B_f=1/3$. Likewise, the strange-to-light pair-production ratio, defined by $u\bar u:d\bar d:s\bar s=1:1:r_s$, is encoded in the measured jet strangeness $\langle S_{\rm flow}\rangle_f \simeq S_f+\frac{3r_s}{2+r_s} \left(\frac{1}{3}- \langle B_{\rm flow}\rangle_f\right)$, predicting a nonzero mean net strangeness even in $u$- and $d$-initiated jets. The conserved-current moments appearing in these relations are independent of the renormalization scale. Their simultaneous measurement therefore provides a direct, flavor-resolved probe of vacuum pair production and quantum-number transport during hadronization.

hep-ph↗

arXiv2026

Transverse Charge Distribution as a Probe of Nucleon Transversity

We introduce the transverse charge distribution as a spin-sensitive charge-flow probe for fragmentation and nucleon tomography. By measuring the angular distribution of net electric charge around a fragmenting quark, this observable relies entirely on the tracking of charged-particle directions and charge signs, strictly bypassing the need for calorimetric energy measurements. At leading twist, the distribution decomposes into an unpolarized charge monopole and a chiral-odd transverse charge dipole. We derive the operator product expansion of these distributions onto charge-weighted collinear moments: a charge monopole and a charge dipole governed by the Collins effect and couples directly to transversity. Applying this formalism to transversely polarized $p^\uparrow p$ collisions at RHIC, we show that charge weighting suppresses the unpolarized monopole background and causes the spin-dependent dipoles from oppositely charged hadrons to add coherently. This coherence strongly enhances the resulting azimuthal asymmetries, establishing a theoretically clean and experimentally precise track-only avenue for extracting transversity.

hep-ph↗

arXiv2026

Nested-GPT for variable-multiplicity parton showers: A case study in the resummation of non-global logarithms

We introduce Nested-GPT, a hierarchical autoregressive Transformer architecture for simulating the variable-multiplicity parton-shower histories. As a controlled benchmark, we study the leading-logarithmic resummation of non-global logarithms in the large-$N_c$ limit, utilizing a stochastic Monte Carlo dipole shower to generate reference training data. We systematically evaluate Nested-GPT against a Transformer flow-matching baseline. The flow-matching framework successfully parameterizes the joint distribution of emission kinematics at fixed multiplicity. Its phase-space representation, however, requires the final number of emissions to be specified externally rather than generated dynamically. Conversely, Nested-GPT strictly enforces the ordered Markovian branching structure, predicting emissions sequentially and dynamically evaluating a learned sequence-termination condition. We benchmark both approaches using gap fraction observables under two complementary training regimes: direct training on vetoed histories and inclusive training followed by an analysis-level veto. The resulting generated samples agree with the reference shower within statistical uncertainties for the observables considered. These results establish Nested-GPT as a physically consistent autoregressive surrogate for variable-multiplicity shower generator and motivate extensions to subleading-logarithmic resummation and finite-$N_c$ color evolution.

hep-ph↗

arXiv2025

SegEarth-R2: Towards Comprehensive Language-guided Segmentation for Remote Sensing Images

Effectively grounding complex language to pixels in remote sensing (RS) images is a critical challenge for applications like disaster response and environmental monitoring. Current models can parse simple, single-target commands but fail when presented with complex geospatial scenarios, e.g., segmenting objects at various granularities, executing multi-target instructions, and interpreting implicit user intent. To drive progress against these failures, we present LaSeRS, the first large-scale dataset built for comprehensive training and evaluation across four critical dimensions of language-guided segmentation: hierarchical granularity, target multiplicity, reasoning requirements, and linguistic variability. By capturing these dimensions, LaSeRS moves beyond simple commands, providing a benchmark for complex geospatial reasoning. This addresses a critical gap: existing datasets oversimplify, leading to sensitivity-prone real-world models. We also propose SegEarth-R2, an MLLM architecture designed for comprehensive language-guided segmentation in RS, which directly confronts these challenges. The model's effectiveness stems from two key improvements: (1) a spatial attention supervision mechanism specifically handles the localization of small objects and their components, and (2) a flexible and efficient segmentation query mechanism that handles both single-target and multi-target scenarios. Experimental results demonstrate that our SegEarth-R2 achieves outstanding performance on LaSeRS and other benchmarks, establishing a powerful baseline for the next generation of geospatial segmentation. All data and code will be released at https://github.com/earth-insights/SegEarth-R2.

cs.CV↗

arXiv2025

ZoomEarth: Active Perception for Ultra-High-Resolution Geospatial Vision-Language Tasks

Ultra-high-resolution (UHR) remote sensing (RS) images offer rich fine-grained information but also present challenges in effective processing. Existing dynamic resolution and token pruning methods are constrained by a passive perception paradigm, suffering from increased redundancy when obtaining finer visual inputs. In this work, we explore a new active perception paradigm that enables models to revisit information-rich regions. First, we present LRS-GRO, a large-scale benchmark dataset tailored for active perception in UHR RS processing, encompassing 17 question types across global, region, and object levels, annotated via a semi-automatic pipeline. Building on LRS-GRO, we propose ZoomEarth, an adaptive cropping-zooming framework with a novel Region-Guided reward that provides fine-grained guidance. Trained via supervised fine-tuning (SFT) and Group Relative Policy Optimization (GRPO), ZoomEarth achieves state-of-the-art performance on LRS-GRO and, in the zero-shot setting, on three public UHR remote sensing benchmarks. Furthermore, ZoomEarth can be seamlessly integrated with downstream models for tasks such as cloud removal, denoising, segmentation, and image editing through simple tool interfaces, demonstrating strong versatility and extensibility.

cs.CV↗

arXiv2025

Biomechanically consistent real-time action recognition for human-robot interaction

This paper presents a novel framework for real-time human action recognition in industrial contexts, using standard 2D cameras. We introduce a complete pipeline for robust and real-time estimation of human joint kinematics, input to a temporally smoothed Transformer-based network, for action recognition. We rely on a new dataset including 11 subjects performing various actions, to evaluate our approach. Unlike most of the literature that relies on joint center positions (JCP) and is offline, ours uses biomechanical prior, eg. joint angles, for fast and robust real-time recognition. Besides, joint angles make the proposed method agnostic to sensor and subject poses as well as to anthropometric differences, and ensure robustness across environments and subjects. Our proposed learning model outperforms the best baseline model, running also in real-time, along various metrics. It achieves 88% accuracy and shows great generalization ability, for subjects not facing the cameras. Finally, we demonstrate the robustness and usefulness of our technique, through an online interaction experiment, with a simulated robot controlled in real-time via the recognized actions.

cs.RO↗

arXiv2025

Accessing nucleon transversity with one-point energy correlators

We propose a novel probe of the nucleon's transversity distribution, $h_1^q$, using the one-point energy correlator (OPEC), an infrared-and-collinear safe jet substructure observable. We demonstrate that in transversely polarized $p^{\uparrow}p$ collisions, the OPEC exhibits a single-spin asymmetry (SSA) with a clean $\sin(\phi_s - \phi_n)$ angular dependence. This method probes SSA over a much wider kinematic range in the angular scale $\theta_n$ compared to traditional measurements of hadron transverse momentum~$j_\perp$, establishing a complementary and systematically distinct channel to study the nucleon's three-dimensional structure at RHIC and the future Electron-Ion Collider.

hep-ph↗

arXiv2025

Revisiting gluon density from the BK equation with kinematical constraint and large x terms

We perform analysis of the small x non-linear evolution equation formulated in momentum space supplemented by higher order terms. The equation is defined in wide range of transverse momentum and longitudinal momentum fraction extending previous studies performed in \cite{Kutak:2003bd,Kutak:2004ym}. The linear part of the equation is motivated by the renormalization group improved small x approach which accounts for resummation of higher orders, and includes collinear splitting function and kinematical constraint. The solution to the equation is then used to perform the fit to Deep Inelastic Scattering reduced cross section data.

hep-ph↗

arXiv2023

Renormalization group improved photon impact factors and the high energy virtual photon scattering

We perform the renormalization group improved collinear resummation of the photon-gluon impact factors. We construct the resummed cross section for virtual photon-photon ($γ^*γ^*$) scattering which incorporates the impact factors and BFKL gluon Green's function up to the next-to-leading logarithmic accuracy in energy. The impact factors include important kinematical effects which are responsible for the most singular poles in Mellin space at next-to-leading order. Further conditions on the resummed cross section are obtained by requiring the consistency with the collinear limits. Our analysis is consistent with previous impact factor calculations at NLO, apart from a new term proportional to $C_F$ that we find for the longitudinal polarization. Finally, we use the resummed cross section to compare with the LEP data on the $γ^*γ^*$ cross section and with previous calculations. The resummed result is lower than the leading logarithmic approximation but higher than the pure next-to-leading one, and is consistent with the experimental data.

hep-ph↗

arXiv2022

Structure functions from renormalization group improved small $x$ evolution

We perform the fit to the structure function $F_2$ data from HERA including terms due to the resummation at small $x$. The equation for the unintegrated gluon density is solved, previously established within the renormalization group improved small $x$ framework. We find very good description of the structure function $F_2$ and its charm component $F_2^c$. The resulting unintegrated gluon density is found to be consistent with the calculations based on similar approaches available in the literature, with only slightly higher intercept value.

hep-ph↗

arXiv2019

On the different forms of the kinematical constraint in BFKL

We perform a detailed analysis of the different forms of the kinematical constraint imposed on the low $x$ evolution that appear in the literature. We find that all of them generate the same leading anti-collinear poles in Mellin space which agree with BFKL up to NLL order and up to NNLL in $N=4$ sYM. The coefficients of subleading poles vanish up to NNLL order for all constraints and we prove that this property should be satisfied to all orders. We then demonstrate that the kinematical constraints differ at further subleading orders of poles. We quantify the differences between the different forms of the constraints by performing numerical analysis both in Mellin space and in momentum space. It can be shown that in all three cases BFKL equation can be recast into the differential form, with the argument of the longitudinal variable shifted by the combination of the transverse coordinates.

hep-ph↗