arXiv subjects

Xukun Wang

Publications and source records attributed to Xukun Wang.

8 recordsLinked to original sources

arXiv2026

Remedying Coarsening-Based GNN Training under Heterophily via Adaptive Complementary Enhancement

Coarsening-based training for graph neural networks (GNNs), i.e.\ training on coarsened graphs rather than the original large ones, has become a promising direction for scaling GNNs to massive graphs. However, prior work has been evaluated almost exclusively on \textit{homophilic} graphs, leaving the more challenging \textit{heterophilic} settings underexplored. We show, both empirically and theoretically, that existing coarsening-based training methods suffer significant performance degradation on heterophilic graphs due to inevitable loss of graph information during coarsening. To address this, we propose {\bf A}daptive {\bf C}omplementary {\bf E}nhancement, a plug-and-play, model-agnostic strategy that reintegrates the information discarded in coarsening: ACE learns a projector for re-constructing original node features and applies \textit{anisotropic structural regularization} to embed local heterophily. We further adopt \textit{homoscedastic uncertainty weighting} to adaptively balance the combined training objective of primary coarsened-graph training loss and full-graph auxiliary loss with augmented node features re-constructed by the heterophily-aware projector. Extensive experiments show that ACE drives consistent gains on heterophilic benchmarks while preserving competitive results on homophilic graphs with minimal computational overhead. Code is available at the GitHub repository: https://github.com/vasile-paskardlgm/ACE.

cs.LG↗

arXiv2026

Projection-Based Reconstruction for Achieving High-Order Accuracy from Low-Order DGSEM Simulations

High-order discontinuous Galerkin spectral element methods (DGSEM) based on Legendre-Gauss-Lobatto (LGL) nodes provide accurate and efficient discretizations for conservation laws. However, their cost, memory footprint, and time-step restrictions increase rapidly when the degree of the polynomial increases. This paper develops a corrected $\mathbb{P}_n\mathbb{P}_m$ ($c\mathbb{P}_n\mathbb{P}_m$) approach for DGSEM-LGL discretizations that aims to recover the accuracy of an $m^{th}$-order approximation while evolving only the degrees of freedom associated with an $n^{th}$-order representation, with $n<m$. The projected evolution of the high-order components is derived first at the continuous level and then in the fully discrete DGSEM-LGL setting. The discrete analysis shows that because LGL quadrature is not exact for the highest Legendre mode, a correction term for that mode is required to preserve the order of convergence. A compact projection-based reconstruction operator is then introduced to recover high-order components without solving the enlarged constrained least-squares systems used in standard reconstruction procedures. For sufficiently smooth solutions, the resulting $c\mathbb{P}_n\mathbb{P}_m$ scheme is shown to achieve the expected $m+1^{th}$ convergence order. Numerical experiments for one- and two-dimensional conservation laws, including Euler, viscous Burgers, and 2D decaying homogeneous isotropic turbulence, confirm theoretical convergence behavior and demonstrate competitive accuracy relative to computational cost, with particularly clear efficiency gains for viscous flows.

math.NA↗

arXiv2026

Noise is not always detrimental: the capacity of quantum batteries is enhanced in black holes

Quantum battery capacity, as a critical metric for quantifying energy storage and release in quantum systems, exhibits complex behaviors in curved spacetime and noisy environments. This study focuses on bipartite mixed state, aiming to explore the modulation of quantum battery capacity by Hawking radiation and environmental noise. We find a counterintuitive phenomenon that Hawking radiation can enhance battery capacity, exerting a positive influence on energy storage, a result that stands in stark contrast to the detrimental effects typically associated with entanglement and coherence. When a quantum battery is simultaneously subjected to environmental noise and Hawking radiation, its capacity generally degrades, with the extent of degradation depending on the type of noise. The charging and discharging behaviors largely follow the same patterns observed in the noiseless scenario; however, under a bit flip channel with strong noise intensity, the charging-discharging pattern reverses. In the extreme case of maximum noise intensity, the capacity of the quantum battery under depolarizing noise tends to zero. The underlying physical mechanism lies in the fact that the bit flip channel disrupts the original population distribution of energy levels, thereby altering the average energy of the system and establishing a perturbative environment for bidirectional energy exchange. This differs fundamentally from the phase flip channel. These findings offer a new perspective for the theory of quantum batteries in noninertial reference frames.

quant-ph↗

arXiv2025

Accelerating high-order energy-stable discontinous Galerkin solver using auto-differentiation and neural networks

High-order Discontinuous Galerkin Spectral Element Methods (DGSEM) provide excellent accuracy for complex flow simulations, but their computational cost increases sharply with higher polynomial orders. %that provide very accurate solutions. To alleviate these limitations, this work presents a differentiable DG solver coupled with neural networks (NNs) that learn corrective forcing terms to correct low-order simulations and provide high-order accuracy. The solver's full differentiability enables gradient-based optimization and interactive (solver-in-the-loop) training, mitigating the data-shift problem typically encountered in static, offline learning. Two representative test cases are considered: the one-dimensional viscous Burgers' equation and two-dimensional decaying homogeneous isotropic turbulence (DHIT). The results demonstrate that interactive training with extended unrolling horizons substantially improves the precision and long-term stability of the simulation compared to static training. For the Burgers' equation, a $\mathbb{P}_2$ simulation corrected using a NN-correction achieves the accuracy of a $\mathbb{P}_4$ solution with eight times reduction in computational cost. For the DHIT case, the NN-corrected low-order simulations successfully achieve high-order accuracy while reduce the error beyond the training interval. These results highlight the potential of differentiable solvers combined with neural networks as a robust and efficient framework for accelerating high-fidelity DG-based fluid simulations.

physics.flu-dyn↗

arXiv2025

Lax functoriality of Hochschild cochain complex

Unlike the Hochschild chain complex of an algebra, the Hochschild cochain complex of an algebra is not functorial. Nonetheless, we show that the Hochschild cochain complex of an algebra even a dg category is of lax functoriality, i.e., there exists a lax functor from bicategory of dg categories to bicategory of $B_\infty$-algebras which sends every dg category to its Hochschild cochain complex. This result is a homotopy version of the lax functoriality of center of an algebra obtained by Davydov, Kong, Runkel, Grady, Oren, et al, in the more general context of dg categories, and extends the restricted functoriality of Hochschild cochain complex of a dg category obtained by Keller to global lax functoriality.

math.KT↗

arXiv2025

A novel parallelizable convergence accelerating method: Pointwise Frequency Damping

This paper proposes a novel class of data-driven acceleration methods for steady-state flow field solvers. The core innovation lies in predicting and assigning the asymptotic limit value for each parameter during iterations based on its own historical data, rather than processing and assigning the entire flow field at once. This approach fundamentally guarantees identical results between serial and parallel computations. Subsequently, a formula for representing the asymptotic limit based on historical data is derived and discretized, yielding a purely algebraic expression.Furthermore, the applicability scope of the method is discussed, along with the underlying reasons for its acceleration capability. A quantitative expression for estimating the speedup ratio is also provided. Extensive validation cases were tested, ranging from the simplest inviscid airfoil flow to complex three-dimensional viscous transonic cruise flow around a aircraft, and solving asymmetric linear systems via GMRES. These tests consistently demonstrate significant acceleration effects with speedup factors ranging from 2.5 to 4. Combined with the near-zero computational overhead of the purely algebraic formulation during the solving process and the inherently parallel-compatible pointwise prediction principle, the results strongly indicate that this method is highly suitable for large-scale industrial mesh computations.

physics.flu-dyn↗

arXiv2025

Lipschitz-Driven Noise Robustness in VQ-AE for High-Frequency Texture Repair in ID-Specific Talking Heads

Audio-driven IDentity-specific Talking Head Generation (ID-specific THG) has shown increasing promise for applications in filmmaking and virtual reality. Existing approaches are generally constructed as end-to-end paradigms, and have achieved significant progress. However, they often struggle to capture high-frequency textures due to limited model capacity. To address these limitations, we adopt a simple yet efficient post-processing framework -- unlike previous studies that focus solely on end-to-end training -- guided by our theoretical insights. Specifically, leveraging the \textit{Lipschitz Continuity Theory} of neural networks, we prove a crucial noise tolerance property for the Vector Quantized AutoEncoder (VQ-AE), and establish the existence of a Noise Robustness Upper Bound (NRoUB). This insight reveals that we can efficiently obtain an identity-specific denoiser by training an identity-specific neural discrete representation, without requiring an extra network. Based on this theoretical foundation, we propose a plug-and-play Space-Optimized VQ-AE (SOVQAE) with enhanced NRoUB to achieve temporally-consistent denoising. For practical deployment, we further introduce a cascade pipeline combining a pretrained Wav2Lip model with SOVQAE to perform ID-specific THG. Our experiments demonstrate that this pipeline achieves \textit{state-of-the-art} performance in video quality and robustness for out-of-distribution lip synchronization, surpassing existing identity-specific THG methods. In addition, the pipeline requires only a couple of consumer GPU hours and runs in real time, which is both efficient and practical for industry applications.

cs.CV↗

arXiv2025

A data-driven convergence booster for accelerating and stabilizing pseudo time-stepping

This paper introduces a novel data-driven convergence booster that not only accelerates convergence but also stabilizes solutions in cases where obtaining a steady-state solution is otherwise challenging. The method constructs a reduced-order model (ROM) of the solution residual using intermediate solutions and periodically solves a least-square problem in the low-dimensional ROM subspace. The second-order approximation of the residual and the use of normal equation distinguish this work from similar approaches in the literature from the methodology perspective. From the application perspective, in contrast to prior studies that focus on linear systems or idealized problems, we rigorously assess the method's performance on realistic computational fluid dynamics (CFD) applications. In addition to reducing the time complexity of point-iterative solvers for linear systems, we demonstrate substantial reductions in the number of pseudo-time steps required for implicit schemes solving the nonlinear Navier-Stokes equations. Across a range of two- and three-dimensional flows-including subsonic inviscid and transonic turbulent cases-the method consistently achieves a 3 to 4 times speedup in wall-clock time. Lastly, the proposed method acts as a robust stabilizer, capable of converging to steady solutions in flows that would otherwise exhibit persistent unsteadiness-such as vortex shedding or transonic buffet-without relying on symmetry boundary conditions.

physics.flu-dyn↗