arXiv subjects

Quan Zheng

Publications and source records attributed to Quan Zheng.

At least 19 recordsLinked to original sources

arXiv2026

APS: Bias-Controlled Adaptive Prototype Simulation for Population-Scale LLM Agents

LLM-agent simulation offers a flexible computational tool for studying population response trajectories that depend on scenario events, memory, demographics, and evolving social context. However, full multi-round simulation scales linearly with both population size and horizon, requiring every agent to query the LLM at every round. We propose Adaptive Prototype Simulation (APS), a framework that reframes scalable LLM-based simulation as a recurrent oracle-allocation problem. APS retains the designated LLM as the online transition oracle while querying adaptive core prototypes, selected singleton-tail agents, and shadow-audit agents. Prototype responses induce local response surfaces for nearby agents, reducing online LLM calls without replacing the underlying transition model. To control approximation bias, shadow-audit residual correction estimates propagation residuals for aggregate correction and future budget allocation, while tail-protected singleton routing directly queries selected isolated, heterogeneous, or high-curvature regions that are vulnerable to smoothing. Theoretically, we treat APS as an estimator for full-scale high-precision individual social simulation and decompose its errors into prototype-coverage error, shadow-audit residual-correction error, local-propagation bias, and temporal context mismatch. Under the reported protocols, APS gives lower reference-aligned distributional discrepancy than scale-oriented and same-budget baselines while reducing online LLM calls, with ablations and compact robustness checks diagnosing the main bias-control mechanisms. In a 10M-agent, multi-round public-opinion simulation, APS achieves a 381.1-fold reduction over full simulation, with reference-aligned final-round JSD of 0.094 against the corresponding full-LLM reference.

cs.MA↗

arXiv2026

Dynamic Mask Enhanced Intelligent Multi-UAV Deployment for Urban Vehicular Networks

Vehicular Ad Hoc Networks (VANETs) play a crucial role in realizing vehicle-road collaboration and intelligent transportation. However, urban VANETs often face challenges such as frequent link disconnections and subnet fragmentation, which hinder reliable connectivity. To address these issues, we dynamically deploy multiple Unmanned Aerial Vehicles (UAVs) as communication relays to enhance VANET. A novel Score based Dynamic Action Mask enhanced QMIX algorithm (Q-SDAM) is proposed for multi-UAV deployment, which maximizes vehicle connectivity while minimizing multi-UAV energy consumption. Specifically, we design a score-based dynamic action mask mechanism to guide UAV agents in exploring large action spaces, accelerate the learning process and enhance optimization performance. The practicality of Q-SDAM is validated using real-world datasets. We show that Q-SDAM improves connectivity by 18.2% while reducing energy consumption by 66.6% compared with existing algorithms.

cs.NI↗

arXiv2026

CLIP-Guided Unsupervised Semantic-Aware Exposure Correction

Improper exposure often leads to severe loss of details, color distortion, and reduced contrast. Exposure correction still faces two critical challenges: (1) the ignorance of object-wise regional semantic information causes the color shift artifacts; (2) real-world exposure images generally have no ground-truth labels, and its labeling entails massive manual editing. To tackle the challenges, we propose a new unsupervised semantic-aware exposure correction network. It contains an adaptive semantic-aware fusion module, which effectively fuses the semantic information extracted from a pre-trained Fast Segment Anything Model into a shared image feature space. Then the fused features are used by our multi-scale residual spatial mamba group to restore the details and adjust the exposure. To avoid manual editing, we propose a pseudo-ground truth generator guided by CLIP, which is fine-tuned to automatically identify exposure situations and instruct the tailored corrections. Also, we leverage the rich priors from the FastSAM and CLIP to develop a semantic-prompt consistency loss to enforce semantic consistency and image-prompt alignment for unsupervised training. Comprehensive experimental results illustrate the effectiveness of our method in correcting real-world exposure images and outperforms state-of-the-art unsupervised methods both numerically and visually.

cs.CV↗

arXiv2026

Fine-grained Verbal Attack Detection via a Hierarchical Divide-and-Conquer Framework

In the digital era, effective identification and analysis of verbal attacks are essential for maintaining online civility and ensuring social security. However, existing research is limited by insufficient modeling of conversational structure and contextual dependency, particularly in Chinese social media where implicit attacks are prevalent. Current attack detection studies often emphasize general semantic understanding while overlooking user response relationships, hindering the identification of implicit and context-dependent attacks. To address these challenges, we present the novel "Hierarchical Attack Comment Detection" dataset and propose a divide-and-conquer, fine-grained framework for verbal attack recognition based on spatiotemporal information. The proposed dataset explicitly encodes hierarchical reply structures and chronological order, capturing complex interaction patterns in multi-turn discussions. Building on this dataset, the framework decomposes attack detection into hierarchical subtasks, where specialized lightweight models handle explicit detection, implicit intent inference, and target identification under constrained context. Extensive experiments on the proposed dataset and benchmark intention detection datasets show that smaller models using our framework significantly outperform larger monolithic models relying on parameter scaling, demonstrating the effectiveness of structured task decomposition.

cs.CL↗

arXiv2025

Pointwise estimates for the fundamental solutions of higher order Schrödinger equations in low odd dimensions

In this paper, we study the fundamental solution of the higher order Schrödinger equation \begin{equation*} \mathrm{i}\partial_t u(x,t) = \big((-Δ)^m + V(x)\big)u(x,t), \quad t \in \mathbb{R}, \ x \in \mathbb{R}^n, \end{equation*} for any odd dimension $n$ and integer $m \geq 1$ satisfying $n < 4m$, where $V$ is a real-valued bounded potential with suitable decay. Let $P_{ac}(H)$ denote the projection onto the absolutely continuous spectral subspace of $H = (-Δ)^m + V$, and assume $H$ has no positive embedded eigenvalues. Our main result says that the evolution operator $e^{-\mathrm{i}tH}P_{ac}(H)$ has an integral kernel $K(t,x,y)$ satisfying the pointwise estimate \begin{equation*} |K(t,x,y)| \leq C (1 + |t|)^{-h} (1 + |t|^{-\frac{n}{2m}}) \left(1 + |t|^{-\frac{1}{2m}}|x - y|\right)^{-\frac{n(m-1)}{2m-1}}, \quad t \neq 0, \ x,y \in \mathbb{R}^n, \end{equation*} where the exponent $h$ depends on $m$, $n$, and the zero energy resonance structure of $H$. We also prove analogous estimates for smoothing operators of the form $H^{\fracα{2m}}e^{-\mathrm{i}tH}P_{ac}(H)$. The key innovation of this paper is a unified approach to deriving asymptotic expansions of the perturbed resolvents around zero, which comprehensively addresses all possible resonance types.

math.AP↗

arXiv2025

Human Locomotion Implicit Modeling Based Real-Time Gait Phase Estimation

Gait phase estimation based on inertial measurement unit (IMU) signals facilitates precise adaptation of exoskeletons to individual gait variations. However, challenges remain in achieving high accuracy and robustness, particularly during periods of terrain changes. To address this, we develop a gait phase estimation neural network based on implicit modeling of human locomotion, which combines temporal convolution for feature extraction with transformer layers for multi-channel information fusion. A channel-wise masked reconstruction pre-training strategy is proposed, which first treats gait phase state vectors and IMU signals as joint observations of human locomotion, thus enhancing model generalization. Experimental results demonstrate that the proposed method outperforms existing baseline approaches, achieving a gait phase RMSE of $2.729 \pm 1.071%$ and phase rate MAE of $0.037 \pm 0.016%$ under stable terrain conditions with a look-back window of 2 seconds, and a phase RMSE of $3.215 \pm 1.303%$ and rate MAE of $0.050 \pm 0.023%$ under terrain transitions. Hardware validation on a hip exoskeleton further confirms that the algorithm can reliably identify gait cycles and key events, adapting to various continuous motion scenarios. This research paves the way for more intelligent and adaptive exoskeleton systems, enabling safer and more efficient human-robot interaction across diverse real-world environments.

cs.RO↗

arXiv2025

SeWA: Selective Weight Average via Probabilistic Masking

Weight averaging has become a standard technique for enhancing model performance. However, methods such as Stochastic Weight Averaging (SWA) and Latest Weight Averaging (LAWA) often require manually designed procedures to sample from the training trajectory, and the results depend heavily on hyperparameter tuning. To minimize human effort, this paper proposes a simple yet efficient algorithm called Selective Weight Averaging (SeWA), which adaptively selects checkpoints during the final stages of training for averaging. Based on SeWA, we show that only a few points are needed to achieve better generalization and faster convergence. Theoretically, solving the discrete subset selection problem is inherently challenging. To address this, we transform it into a continuous probabilistic optimization framework and employ the Gumbel-Softmax estimator to learn the non-differentiable mask for each checkpoint. Further, we theoretically derive the SeWA's stability-based generalization bounds, which are sharper than that of SGD under both convex and non-convex assumptions. Finally, solid extended experiments in various domains, including behavior cloning, image classification, and text classification, further validate the effectiveness of our approach.

cs.LG↗

arXiv2024

Pointwise estimates for the fundamental solutions of higher order Schrödinger equations in odd dimensions II: high dimensional case

In this paper, for any odd $n$ and any integer $m\geq1$ with $n>4m$, we study the fundamental solution of the higher order Schrödinger equation \begin{equation*} \mathrm{i}\partial_tu(x,t)=((-Δ)^m+V(x))u(x,t),\quad t\in \mathbb{R},\,\,x\in \mathbb{R}^n, \end{equation*} where $V$ is a real-valued $C^{\frac{n+1}{2}-2m}$ potential with certain decay. Let $P_{ac}(H)$ denote the projection onto the absolutely continuous spectrum space of $H=(-Δ)^m+V$, and assume that $H$ has no positive embedded eigenvalue. Our main result says that $e^{-\mathrm{i}tH}P_{ac}(H)$ has integral kernel $K(t,x,y)$ satisfying \begin{equation*} |K(t, x,y)|\le C(1+|t|)^{-(\frac{n}{2m}-σ)}(1+|t|^{-\frac{n}{2 m}})\left(1+|t|^{-\frac{1}{2 m}}|x-y|\right)^{-\frac{n(m-1)}{2 m-1}},\quad t\neq0,\,x,y\in\mathbb{R}^n, \end{equation*} where $σ=2$ if $0$ is an eigenvalue of $H$, and $σ=0$ otherwise. A similar result for smoothing operators $H^\fracα{2m}e^{-\mathrm{i}tH}P_{ac}(H)$ is also given. The regularity condition $V\in C^{\frac{n+1}{2}-2m}$ is optimal in the second order case, and it also seems optimal when $m>1$.

math.AP↗

arXiv2024

Learning to Importance Sample in Primary Sample Space

Importance sampling is one of the most widely used variance reduction strategies in Monte Carlo rendering. In this paper, we propose a novel importance sampling technique that uses a neural network to learn how to sample from a desired density represented by a set of samples. Our approach considers an existing Monte Carlo rendering algorithm as a black box. During a scene-dependent training phase, we learn to generate samples with a desired density in the primary sample space of the rendering algorithm using maximum likelihood estimation. We leverage a recent neural network architecture that was designed to represent real-valued non-volume preserving ('Real NVP') transformations in high dimensional spaces. We use Real NVP to non-linearly warp primary sample space and obtain desired densities. In addition, Real NVP efficiently computes the determinant of the Jacobian of the warp, which is required to implement the change of integration variables implied by the warp. A main advantage of our approach is that it is agnostic of underlying light transport effects, and can be combined with many existing rendering techniques by treating them as a black box. We show that our approach leads to effective variance reduction in several practical scenarios.

cs.LG↗

arXiv2024

Learning Novel View Synthesis from Heterogeneous Low-light Captures

Neural radiance field has achieved fundamental success in novel view synthesis from input views with the same brightness level captured under fixed normal lighting. Unfortunately, synthesizing novel views remains to be a challenge for input views with heterogeneous brightness level captured under low-light condition. The condition is pretty common in the real world. It causes low-contrast images where details are concealed in the darkness and camera sensor noise significantly degrades the image quality. To tackle this problem, we propose to learn to decompose illumination, reflectance, and noise from input views according to that reflectance remains invariant across heterogeneous views. To cope with heterogeneous brightness and noise levels across multi-views, we learn an illumination embedding and optimize a noise map individually for each view. To allow intuitive editing of the illumination, we design an illumination adjustment module to enable either brightening or darkening of the illumination component. Comprehensive experiments demonstrate that this approach enables effective intrinsic decomposition for low-light multi-view noisy images and achieves superior visual quality and numerical performance for synthesizing novel views compared to state-of-the-art methods.

cs.CV↗

arXiv2024

3D-SceneDreamer: Text-Driven 3D-Consistent Scene Generation

Text-driven 3D scene generation techniques have made rapid progress in recent years. Their success is mainly attributed to using existing generative models to iteratively perform image warping and inpainting to generate 3D scenes. However, these methods heavily rely on the outputs of existing models, leading to error accumulation in geometry and appearance that prevent the models from being used in various scenarios (e.g., outdoor and unreal scenarios). To address this limitation, we generatively refine the newly generated local views by querying and aggregating global 3D information, and then progressively generate the 3D scene. Specifically, we employ a tri-plane features-based NeRF as a unified representation of the 3D scene to constrain global 3D consistency, and propose a generative refinement network to synthesize new contents with higher quality by exploiting the natural image prior from 2D diffusion model as well as the global 3D information of the current scene. Our extensive experiments demonstrate that, in comparison to previous methods, our approach supports wide variety of scene generation and arbitrary camera trajectories with improved visual quality and 3D consistency.

cs.CV↗

arXiv2023

Neural Invertible Variable-degree Optical Aberrations Correction

Optical aberrations of optical systems cause significant degradation of imaging quality. Aberration correction by sophisticated lens designs and special glass materials generally incurs high cost of manufacturing and the increase in the weight of optical systems, thus recent work has shifted to aberration correction with deep learning-based post-processing. Though real-world optical aberrations vary in degree, existing methods cannot eliminate variable-degree aberrations well, especially for the severe degrees of degradation. Also, previous methods use a single feed-forward neural network and suffer from information loss in the output. To address the issues, we propose a novel aberration correction method with an invertible architecture by leveraging its information-lossless property. Within the architecture, we develop conditional invertible blocks to allow the processing of aberrations with variable degrees. Our method is evaluated on both a synthetic dataset from physics-based imaging simulation and a real captured dataset. Quantitative and qualitative experimental results demonstrate that our method outperforms compared methods in correcting variable-degree optical aberrations.

cs.CV↗

arXiv2023

A survey on facial image deblurring

When a facial image is blurred, it significantly affects high-level vision tasks such as face recognition. The purpose of facial image deblurring is to recover a clear image from a blurry input image, which can improve the recognition accuracy, etc. However, general deblurring methods do not perform well on facial images. Therefore, some face deblurring methods have been proposed to improve performance by adding semantic or structural information as specific priors according to the characteristics of the facial images. In this paper, we survey and summarize recently published methods for facial image deblurring, most of which are based on deep learning. First, we provide a brief introduction to the modeling of image blurring. Next, we summarize face deblurring methods into two categories: model-based methods and deep learning-based methods. Furthermore, we summarize the datasets, loss functions, and performance evaluation metrics commonly used in the neural network training process. We show the performance of classical methods on these datasets and metrics and provide a brief discussion on the differences between model-based and learning-based methods. Finally, we discuss the current challenges and possible future research directions.

cs.CV↗

arXiv2023

Macdonald-Hurwitz Number

Inspired by J. Novak's works on the asymptotic behavior of the BGW and the HCIZ matrix integrals \cite{[N0]} and by the algebraic and geometric properties of the Hurwitz numbers \cite{[IP]}, \cite{[LZZ]}, \cite{[LR]}, \cite{[OP]}, \cite{[Z1]}, and by the symplectic surgery theory of the relative GW-invariants \cite{[IP]}, \cite{[LR]}, using the elements of the transform matrix from the integral Macdonald function with two parameters to the homogeneous symmetric power sum functions \cite{[M]}, we have constructed the Macdonald-Hurwitz numbers. As an application, we have constructed a series of new genus-expanded cut-and-join differential operators, which can be thought of as the generalization of the Laplace-Beltrami operators and have the genus-expanded integral Macdonald functions as their common eigenfunctions. We have also obtained some generating wave functions of the same degree, which are generated by the Macdonald-Hurwitz numbers and can be expressed in terms of the new cut-and-join differential operators and the initial values. Another application is that we have constructed a new commutative associative algebra $(C(\mathbb{F}[S_{d}]),\circ_{q,t})$ (referring to the last section (6)). By taking the limit along a special path $η(A|B)$ (referring to the formulas (140), (141)), we specialize $(C(\mathbb{F}[S_{d}]),\circ_{q,t})$ to be a commutative associative algebra $(C(\hat{\mathbb{F}}[S_{d}]),\circ_{A|B})$, which will be proven to be isomorphic to the middle-dimensional $\mathbb{\mathbb{C}^*}$-equivalent cohomological rings via the Jack functions over the Hilbert scheme points of $\mathbb{C}^2$ constructed by W. Li, Z. Qin, and W. Wang in \cite{[LQW2]}.}

math.SG↗

arXiv2022

Genus Expanded Cut-and-Join operators and generalized Hurwtiz numbers

To distinguish the contributions to the generalized Hurwitz number of the source Riemann surface with different genus, we define the genus expanded cut-and-join operators by observing carefully the symplectic surgery and the gluing formulas of the relative GW-invariants. As an application, we get some differential equations for the generating functions of the generalized Hurwitz numbers for the source Riemann surface with different genus, thus we can express the generating functions in terms of the genus expanded cut-and-join operators.

math.SG↗

arXiv2022

Shifted genus expanded $\cal{W}_{\infty}$ algebra and shifted Hurwtiz numbers

We construct the shifted genus expanded $\cal{W}_{\infty}$ algebra, which is isomorphic to the central subalgebra $\cal{A}_{\infty}$ of infinite symmetric group algebra and to the shifted Schur symmetrical function algebra $Λ^\ast$ defined by A. Y. Okounkov and G. I. Olshanskii. As an application, we get some differential equations for the generating functions of the shifted Hurwitz numbers, thus we can express the generating functions in terms of the shifted genus expanded cut-and-join operators.

math.SG↗

arXiv2022

Uniform Complex Time Heat Kernel Estimates Without Gaussian Bounds

In this paper, first we consider the uniform complex time heat kernel estimates of $e^{-z(-Δ)^{\fracα{2}}}$ for $α>0, z\in \mathbb{C}^+$. When $\fracα{2}$ is not an integer, generally the heat kernel doest not have the Gaussian upper bounds for real time. Thus the Phragmén-Lindelöf methods fail to give the uniform complex time estimates. Instead, our first result gives the asymptotic estimates for $P(z, x)$ as $z$ tending to the imaginary axis. Then we prove the uniform complex time heat kernel estimates. Finally we also show the uniform estimates of analytic semigroup generated by $H=(-Δ)^{\fracα{2}}+V$ where $V$ belongs to higher order Kato class.

math.CA↗

arXiv2022

Dispersive estimates for the Schrödinger equation with finite rank perturbations

In this paper, we investigate dispersive estimates for the time evolution of Hamiltonians $$ H=-Δ+\sum_{j=1}^N\langle\cdot\,, φ_j\rangle φ_j\quad\,\,\,\text{in}\,\,\,\mathbb{R}^d,\,\, d\ge 1, $$ where each $φ_j$ satisfies certain smoothness and decay conditions. We show that, under a spectral assumption, there exists a constant $C=C(N, d, φ_1,\ldots, φ_N)>0$ such that $$ \|e^{-itH}\|_{L^1-L^{\infty}}\leq C t^{-\frac{d}{2}}, \,\,\,\text{for}\,\,\, t>0. $$ As far as we are aware, this seems to provide the first study of $L^1-L^{\infty}$ estimates for finite rank perturbations of the Laplacian in any dimension. We first deal with rank one perturbations ($N=1$). Then we turn to the general case. The new idea in our approach is to establish the Aronszajn-Krein type formula for finite rank perturbations. This allows us to reduce the analysis to the rank one case and solve the problem in a unified manner. Moreover, we show that in some specific situations, the constant $C(N, d, φ_1,\ldots, φ_N)$ grows polynomially in $N$. Finally, as an application, we are able to extend the results to $N=\infty$ and deal with some trace class perturbations.

math.AP↗