arXiv subjects

Hong Qin

Publications and source records attributed to Hong Qin.

At least 37 records · Page 2Linked to original sources

arXiv2025

Closed-form Solutions: A New Perspective on Solving Differential Equations

The quest for analytical solutions to differential equations has traditionally been constrained by the need for extensive mathematical expertise. Machine learning methods like genetic algorithms have shown promise in this domain, but are hindered by significant computational time and the complexity of their derived solutions. This paper introduces SSDE (Symbolic Solver for Differential Equations), a novel reinforcement learning-based approach that derives symbolic closed-form solutions for various differential equations. Evaluations across a diverse set of ordinary and partial differential equations demonstrate that SSDE outperforms existing machine learning methods, delivering superior accuracy and efficiency in obtaining analytical solutions.

cs.LG↗

arXiv2025

Spin-weighted spherical harmonics as massless angular momentum eigenstates and their role in obstructing spin-orbital decompositions

We show that for massless helicity $h$ particles, the angular momentum eigenstates are given in an appropriate coordinate system by the spin-weighted spherical harmonics ${_{-h}Y_{jm}}$ of spin-weight $-h$. In particular, these are simultaneous eigenstates of the Hamiltonian, helicity, $J^2$, and $J_z$. The appearance of the spin-weighted spherical harmonics as opposed to the ordinary spherical harmonics reflects the nontrivial topological structure of massless particles with nonzero helicity. The resultant angular momentum multiplet structure is quite different than that of massive particles, with at most one multiplet for each angular momentum $j$ and with $|h|$ acting as a lower bound on $j$. This illustrates the obstruction to a spin-orbital decomposition of the angular momentum for massless particles, as such a sparse multiplet structure is not consistent with any reasonable spin-orbital splitting.

math-ph↗

arXiv2025

The topology, geometry, and angular momentum of cold plasma waves

It was recently discovered that plasma waves possess topologically protected edge modes, indicating the existence of topologically nontrivial structures in the governing equations. Here we give a rigorous study of the underlying topological vector bundle structure of cold unmagnetized plasma waves and show that this topology can be used to uncover a number of new results about these waves. The topological properties of the electromagnetic waves mirror those recently found for photons and other massless particles. We show that there exists an explicit globally smooth polarization basis for electromagnetic plasma waves -- surprisingly, this does not violate the hairy ball theorem. The rotational symmetry of the waves gives a natural decomposition into topologically nontrivial $R$ and $L$ circularly polarized electromagnetic waves and the topologically trivial electrostatic Langmuir waves. The existence of topologically nontrivial waves, despite the effective mass introduced by the plasma, is related to the resonance of electrostatic and electromagnetic waves. We show that the eigenstates of the angular momentum operator are the spin-weighted spherical harmonics, giving a novel globally smooth basis for plasma waves. The sparseness of the resultant angular momentum multiplet structure illustrates that the angular momentum does not split into well-defined spin and orbital parts. However, we demonstrate that the angular momentum admits a natural decomposition, induced by the rotational symmetry, into two quasi-angular momentum components, termed helicity and orbital quasi-angular momentum. Although these operators do not generate physical rotations and therefore do not qualify as true angular momentum operators, they are gauge invariant, well-defined, and appear to be experimentally relevant.

physics.plasm-ph↗

arXiv2025

Second quantization of nonlinear Vlasov-Poisson system for quantum computation

The Vlasov-Poisson equations, fundamental in plasma physics and astrophysical applications, are rendered linear, finite-dimensional, and discrete by second quantization. Conditions for correspondence between the pre-quantized and quantized equations are derived, and numerical simulations demonstrating the quantized linear system can capture nonlinear dynamics are presented. Finally, encouraging scaling relations emphasizing the prospect of using quantum computers to efficiently integrate the second quantized Vlasov-Poisson equations as a model for the usual Vlasov-Poisson equations are derived.

physics.plasm-ph↗

arXiv2025

Gap modes in Arnold tongues and their topological origins

Gap modes in a modified Mathieu equation, perturbed by a Dirac delta potential, are investigated. It is proved that the modified Mathieu equation admits stable isolated gap modes with topological origins in the unstable regions of the Mathieu equation, which are known as Arnold tongues. The modes may be identified as localized electron wavefunctions in a 1D chain or as toroidal Alfvén eigenmodes. A generalization of this argument shows that gap modes can be induced in regimes of instability by localized potential perturbations for a large class of periodic Hamiltonians.

math-ph↗

arXiv2025

GeoMaNO: Geometric Mamba Neural Operator for Partial Differential Equations

The neural operator (NO) framework has emerged as a powerful tool for solving partial differential equations (PDEs). Recent NOs are dominated by the Transformer architecture, which offers NOs the capability to capture long-range dependencies in PDE dynamics. However, existing Transformer-based NOs suffer from quadratic complexity, lack geometric rigor, and thus suffer from sub-optimal performance on regular grids. As a remedy, we propose the Geometric Mamba Neural Operator (GeoMaNO) framework, which empowers NOs with Mamba's modeling capability, linear complexity, plus geometric rigor. We evaluate GeoMaNO's performance on multiple standard and popularly employed PDE benchmarks, spanning from Darcy flow problems to Navier-Stokes problems. GeoMaNO improves existing baselines in solution operator approximation by as much as 58.9%.

cs.LG↗

arXiv2025

Improved Dimensionality Reduction for Inverse Problems in Nuclear Fusion and High-Energy Astrophysics

Many inverse problems in nuclear fusion and high-energy astrophysics research, such as the optimization of tokamak reactor geometries or the inference of black hole parameters from interferometric images, necessitate high-dimensional parameter scans and large ensembles of simulations to be performed. Such inverse problems typically involve large uncertainties, both in the measurement parameters being inverted and in the underlying physics models themselves. Monte Carlo sampling, when combined with modern non-linear dimensionality reduction techniques such as autoencoders and manifold learning, can be used to reduce the size of the parameter spaces considerably. However, there is no guarantee that the resulting combinations of parameters will be physically valid, or even mathematically consistent. In this position paper, we advocate adopting a hybrid approach that leverages our recent advances in the development of formal verification methods for numerical algorithms, with the goal of constructing parameter space restrictions with provable mathematical and physical correctness properties, whilst nevertheless respecting both experimental uncertainties and uncertainties in the underlying physical processes.

cs.LG↗

arXiv2025

Zero curvature is a necessary and sufficient condition for a spin-orbital decomposition

There has been an extended debate regarding the existence of a spin-orbital decomposition of the angular momentum of photons and other massless particles. It was recently shown that there are both geometric and topological obstructions preventing any such decomposition. Here we show that any geometric connection on a particle's state space induces a splitting of the angular momentum into two operators. These operators are well-defined angular momentum operators if and only if the connection has zero curvature. Massive particles have two canonical curved connections corresponding to boosts and rotations, respectively. These can be uniquely combined to produce a flat connection, and this gives a novel derivation of the Newton-Wigner position operator and the corresponding spin and orbital angular momenta for relativistic massive particles. When the mass is taken to zero, transverse boosts and rotations degenerate, leaving only a single connection for massless particles. This connection produces a commonly proposed splitting of the massless angular momentum into two operators. However, the connection is not flat, explaining why these operators do not satisfy the angular momentum commutation relations and are thus not true spin and orbital angular momentum operators.

math-ph↗

arXiv2025

Details Enhancement in Unsigned Distance Field Learning for High-fidelity 3D Surface Reconstruction

While Signed Distance Fields (SDF) are well-established for modeling watertight surfaces, Unsigned Distance Fields (UDF) broaden the scope to include open surfaces and models with complex inner structures. Despite their flexibility, UDFs encounter significant challenges in high-fidelity 3D reconstruction, such as non-differentiability at the zero level set, difficulty in achieving the exact zero value, numerous local minima, vanishing gradients, and oscillating gradient directions near the zero level set. To address these challenges, we propose Details Enhanced UDF (DEUDF) learning that integrates normal alignment and the SIREN network for capturing fine geometric details, adaptively weighted Eikonal constraints to address vanishing gradients near the target surface, unconditioned MLP-based UDF representation to relax non-negativity constraints, and DCUDF for extracting the local minimal average distance surface. These strategies collectively stabilize the learning process from unoriented point clouds and enhance the accuracy of UDFs. Our computational results demonstrate that DEUDF outperforms existing UDF learning methods in both accuracy and the quality of reconstructed surfaces. Our source code is at https://github.com/GiliAI/DEUDF.

cs.CV↗

arXiv2025

On Minimizing Phase Space Energies

A primary technical challenge for harnessing fusion energy is to control and extract energy from a non-thermal distribution of charged particles. The fact that phase space evolves by symplectomorphisms fundamentally limits how a distribution may be manipulated. While the constraint of phase-space volume preservation is well understood, other constraints remain to be fully appreciated. To better understand these constraints, we study the problem of extracting energy from a distribution of particles using area-preserving and symplectic linear maps. When a quadratic potential is imposed, we find that the maximal extractable energy can be computed as trace minimization problems. We solve these problems and show that the extractable energy under linear symplectomorphisms may be much smaller than the extractable energy under special linear maps. The method introduced in the present study enables an energy-based proof of the linear Gromov non-squeezing theorem.

math-ph↗

arXiv2025

Energy-momentum-conserving stochastic differential equations and algorithms for nonlinear Landau-Fokker-Planck equation

Coulomb collision is a fundamental diffusion process in plasmas that can be described by the Landau-Fokker-Planck (LFP) equation or the stochastic differential equation (SDE). While energy and momentum are conserved exactly in the LFP equation, they are conserved only on average by the conventional corresponding SDEs, suggesting that the underlying stochastic process may not be well-defined by such SDEs. In this study, we derive new SDEs with exact energy-momentum conservation for the Coulomb collision by factorizing the collective effect of field particles into individual particles and enforcing Newton's third law. These SDEs, when interpreted in the Stratonovich sense, have a particularly simple form that represents pure diffusion between particles without drag. Numerical algorithms that preserve discrete conservation laws are developed and benchmarked in various relaxation processes. Techniques to reduce computational complexity are also discussed.

physics.plasm-ph↗

arXiv2025

Four no-go theorems on the existence of spin and orbital angular momentum of massless bosons

The past decades have seen substantial interest in the so-called orbital angular momentum (OAM) of light, driven largely by its diverse range of applications. However, there are fundamental theoretical issues with decomposing the angular momentum of massless particles, such as photons, into spin (SAM) and orbital angular momentum parts. While the angular momentum of massive particles has a natural splitting into the Wigner SAM and OAM, there are numerous proposed splittings for photons and no consensus about which is correct. Moreover, it has been shown that most of the proposed SAM and OAM operators do not satisfy the defining commutation relations of angular momentum operators and are thus not legitimate splittings. Here, we prove that it is generally impossible to split the total angular momentum operator of massless bosons, such as photons and gravitons, into spin and orbital parts. We prove two further generalizations of this result, showing that there are no SAM-OAM splittings even if (1) the SAM operator generates non-internal symmetries or (2) if one allows the SAM and OAM operators to generate non-SO(3) symmetries.

math-ph↗

arXiv2025

From Transparent to Opaque: Rethinking Neural Implicit Surfaces with $α$-NeuS

Traditional 3D shape reconstruction techniques from multi-view images, such as structure from motion and multi-view stereo, face challenges in reconstructing transparent objects. Recent advances in neural radiance fields and its variants primarily address opaque or transparent objects, encountering difficulties to reconstruct both transparent and opaque objects simultaneously. This paper introduces $α$-Neus -- an extension of NeuS -- that proves NeuS is unbiased for materials from fully transparent to fully opaque. We find that transparent and opaque surfaces align with the non-negative local minima and the zero iso-surface, respectively, in the learned distance field of NeuS. Traditional iso-surfacing extraction algorithms, such as marching cubes, which rely on fixed iso-values, are ill-suited for such data. We develop a method to extract the transparent and opaque surface simultaneously based on DCUDF. To validate our approach, we construct a benchmark that includes both real-world and synthetic scenes, demonstrating its practical utility and effectiveness. Our data and code are publicly available at https://github.com/728388808/alpha-NeuS.

cs.CV↗

arXiv2025

Oblique photons, plasmons, and current-plasmons in relativistic plasmas and their topological implications

Photons in vacuum are transverse in any inertial frame; longitudinal photons only exist virtually. By developing a manifestly covariant theory for electromagnetic excitations in relativistic plasmas and applying Wigner's little group method for elementary particle classifications, we show that photons in plasmas are neither transverse nor longitudinal; they are oblique. Plasmons are electromagnetic and oblique as well. The Lorentz invariant characteristics that distinguishes photons and plasmons is covariant compressibility. The manifestly covariant theory predicts the existence of the current-plasmon, a third oblique, electromagnetic eigenmode, and it also enables the study of photon topology in plasmas. Plasmas remove the photon's Dirac point in vacuum by giving it an effective mass, but create a tilted Dirac-Weyl point by reviving the virtual longitudinal photon. The manifest covariance of the theory demonstrates that relativistic transparency, despite being widely studied, does not exist in plasmas.

physics.plasm-ph↗

arXiv2024

Explainable convolutional neural network model provides an alternative genome-wide association perspective on mutations in SARS-CoV-2

Identifying mutations of SARS-CoV-2 strains associated with their phenotypic changes is critical for pandemic prediction and prevention. We compared an explainable convolutional neural network (CNN) approach and the traditional genome-wide association study (GWAS) on the mutations associated with WHO labels of SARS-CoV-2, a proxy for virulence phenotypes. We trained a CNN classification model that can predict genomic sequences into Variants of Concern (VOCs) and then applied Shapley Additive explanations (SHAP) model to identify mutations that are important for the correct predictions. For comparison, we performed traditional GWAS to identify mutations associated with VOCs. Comparison of the two approaches shows that the explainable neural network approach can more effectively reveal known nucleotide substitutions associated with VOCs, such as those in the spike gene regions. Our results suggest that explainable neural networks for genomic sequences offer a promising alternative to the traditional genome wide analysis approaches.

q-bio.GN↗

arXiv2024

Gromov ground state in phase space engineering for fusion energy

Phase space engineering by RF waves plays important roles in both thermal D-T fusion and non-thermal advanced fuel fusion. But not all phase space manipulation is allowed, certain fundamental limits exist. In addition to Liouville's theorem, which requires the manipulation to be volume-preserving, Gromov's non-squeezing theorem imposes another constraint. The Gardner ground state is defined as the ground state accessible by smooth volume-preserving maps. However, the extra Gromov constraint should produce a higher-energy ground state. An example of a Gardner ground state forbidden by Gromov's non-squeezing theorem is given. The challenge question is: What is the Gromov ground state, i.e., the lowest energy state accessible by smooth symplectic maps? This is a difficult problem. As a simplification, we conjecture that the linear Gromov ground state problem is solvable.

physics.plasm-ph↗

arXiv2024

Graviton topology

Over the past three decades, it has been shown that discrete and continuous media can support topologically nontrivial waves. Recently, it was shown that the same is true of the vacuum, in particular, right (R) and left (L) circularly polarized photons are topologically nontrivial. Here, we study the topology of another class of massless particles, namely gravitons. We show that the collection of all gravitons forms a topologically trivial vector bundle over the lightcone, allowing us to construct a globally smooth basis for gravitons. The graviton bundle also has a natural geometric splitting into two topologically nontrivial subbundles, consisting of the R and L gravitons. The R and L gravitons are unitary irreducible bundle representations of the Poincaré group, and are thus elementary particles; their topology is characterized by the Chern numbers $\mp 4$. This nontrivial topology obstructs the splitting of graviton angular momentum into spin and orbital angular momentum.

math-ph↗

arXiv2024

UGrid: An Efficient-And-Rigorous Neural Multigrid Solver for Linear PDEs

Numerical solvers of Partial Differential Equations (PDEs) are of fundamental significance to science and engineering. To date, the historical reliance on legacy techniques has circumscribed possible integration of big data knowledge and exhibits sub-optimal efficiency for certain PDE formulations, while data-driven neural methods typically lack mathematical guarantee of convergence and correctness. This paper articulates a mathematically rigorous neural solver for linear PDEs. The proposed UGrid solver, built upon the principled integration of U-Net and MultiGrid, manifests a mathematically rigorous proof of both convergence and correctness, and showcases high numerical accuracy, as well as strong generalization power to various input geometry/values and multiple PDE formulations. In addition, we devise a new residual loss metric, which enables unsupervised training and affords more stability and a larger solution space over the legacy losses.

math.NA↗