SearcharxivSearch

arXiv subjects

Feng Guo

Publications and source records attributed to Feng Guo.

At least 19 recordsLinked to original sources

The Local Embedding Problem for Hardy Spaces of Dirichlet Series

We solve the local embedding problem for Hardy spaces of Dirichlet series, which is a dimension-free trace problem asking whether the global $\mathscr{H}^p$-norm controls local $L^p$-mass on the critical line $\operatorname{Re}s=1/2$. More precisely, for every $2<p<\infty$, there exists a constant $C_p<\infty$ such that every Dirichlet polynomial $P$ satisfies $$ \sup_{\theta\in\mathbb{R}}\int_{\theta}^{\theta+1}\left|P\left(\frac12+it\right)\right|^p\,\mathrm{d}t\le C_p\left\lVert P\right\rVert_{\mathscr{H}^p}^{p}, $$ with $C_p$ independent of the number and choice of prime variables on which $P$ depends. Before the present work, the embedding was known at $p=2$ and, by taking integer powers, at the even exponents $p=2k$; it had been conjectured that these exhaust the finite positive cases above $2$. Together with the known failure for $0<p<2$, our theorem gives the sharp finite-exponent classification: the local embedding property holds exactly for $p\ge2$. Thus the true threshold is $p=2$, rather than even integrality. The proof passes to the dual exponent $q=p/(p-1)\in(1,2)$, where an exact frequency decomposition isolates a single resonant Euler-product term. A covariance-preserving replacement of the shared prime factors reduces the resulting moment estimate to a log-correlated Gaussian field, and a critical branching-random-walk bound supplies the required multiscale decay. A finite-cyclic square-function estimate assembles the resonant scales, and Hardy-quotient duality converts the resulting vector-valued bound into the critical-line trace. For $1\le p<\infty$, known equivalences give the same sharp threshold in several classical problems, including the conformally invariant half-plane embedding, the reverse local Carleson-measure transfer, and boundedness of all characteristic-zero Gordon--Hedenmalm composition operators.

math.FA

A Functional Central Limit Theorem for Window Counts of Hardy--Szeg\H{o} Zeros

The Hardy--Szeg\H{o} zero process, investigated in the disk by Peres and Vir\'ag through the independent identically distributed Gaussian analytic function, is a canonical conformally invariant determinantal point process. This paper studies its upper half-plane realization. Although conformally equivalent to the disk model, this realization has its own natural geometry: real-translation invariance turns the process into a stationary object along the boundary and makes long horizontal windows the natural observables. For every admissible height window, we prove a Donsker-type functional central limit theorem for the centered zero counts in expanding horizontal windows, with an explicit intensity and variance depending on the height window. The proof is based on factorial cumulants and Brillinger mixing. The main technical input is a family of all-order integrability estimates for reduced cumulant densities, obtained by exploiting the determinantal cycle structure before integrating over the height variables. As further consequences, we derive an explicit covariance density, an asymptotic variance formula, and a macroscopic Gaussian white-noise limit for linear statistics.

math.PR

Critical Gaussian Multiplicative Chaos on the Circle Is Rajchman

We prove that the Fourier coefficients of the canonical critical Gaussian multiplicative chaos on the circle vanish almost surely at infinity. More precisely, let $M_\phi^{\mathrm{crit}}$ be the canonical critical chaos associated with the centered circle field $\phi$ of covariance $\mathbb{E}[\phi(\theta)\phi(\theta')] = \log\frac{1}{\lvert e^{i\theta}-e^{i\theta'}\rvert}$. Then, almost surely, $\widehat{M_\phi^{\mathrm{crit}}}(n)\longrightarrow0$ as $\lvert n\rvert\to\infty$. This resolves the almost-sure critical Rajchman problem for the canonical circle field. Since critical chaos has Fourier dimension zero almost surely, no positive polynomial Fourier-decay rate can hold; the theorem therefore exhibits qualitative Fourier cancellation beyond the regime of positive Fourier dimension. The proof addresses two coupled difficulties: the heavy, nonuniform cell masses of critical chaos and the need to control exponentially many frequencies in each dyadic annulus. For an auxiliary periodized compact-range star-scale field, a derivative-rooted Bessel regression yields weighted small-cell summability and moving-tail control of exceptional large cells. After conditioning at a coarse scale below the Fourier scale, finite-range independence and conditional Bernstein concentration reduce uniform control of the terminal Fourier coefficients over each dyadic annulus to a spatial-variation estimate for a coarse predictable measure. A smooth positive-definite covariance correction and critical-chaos uniqueness then transfer the Rajchman property to the canonical critical chaos of the exact circle field.

math.PR

Critical Norm Profiles for Finite-Prime Composition Operators on the Hardy Space of Dirichlet Series

We identify the critical boundary operator-norm profile of finite-prime composition operators on the Hardy--Hilbert space \(\mathcal H^2\) of Dirichlet series. For \[ \varphi_{\delta,\boldsymbol\rho}(s) = \frac12+\delta + \delta\sum_{j=1}^d\rho_jp_j^{-s}, \qquad \boldsymbol\rho\in B_d, \] the renormalized positive coefficient operators converge uniformly in operator norm, with \(O(\delta)\) error, to an explicit multivariate weighted Hankel operator \(\mathcal H_{\boldsymbol\rho}\); consequently, \[ 2\delta\|C_{\varphi_{\delta,\boldsymbol\rho}}\|^2 = \|\mathcal H_{\boldsymbol\rho}\| + O(\delta) \] uniformly over \(B_d\). We show that the limiting operator admits the total-degree reduction \[ \mathcal H_{\boldsymbol\rho} \simeq D_{\boldsymbol\rho} H_{R_{\boldsymbol\rho}/2} D_{\boldsymbol\rho}\oplus\mathbf{0}, \] where the diagonal factors are convolution-collision norms of the normalized prime weights. This structure, together with the affine comparison principle of Brevig and Perfekt, yields an explicit concentration inequality for \(\|\mathcal H_{\boldsymbol\rho}\|\), identifies the one-prime configurations as the exact equality cases in the limiting norm estimate, and gives a quantitative deficit away from them. For fixed \(\sigma>\frac12\), we also obtain a second-order expansion of the squared norm and fully finite-dimensional approximations with explicit total-degree and Dirichlet-sum truncation errors. Together, these results show that a single coefficient-operator structure governs the singular boundary profile, the fixed-\(\sigma\) perturbative regime, and certified finite-dimensional approximation.

math.FA

Two Regularity Problems on Analytic Tent Spaces

We study two regularity problems on Hardy-type analytic tent spaces $\mathcal{AT}^p_{q,\alpha}$ on the unit disk: fractional integration and randomization of Taylor coefficients. For fractional integration, we characterize completely the boundedness and compactness of the Hadamard, Flett, and Riemann--Liouville operators between analytic tent spaces, and obtain parallel results for analytic Triebel--Lizorkin spaces. In particular, the case $t=0$ yields a complete solution to the corresponding embedding problem for analytic tent spaces. For randomization, we characterize completely when the random Taylor series $\mathcal{R}f$ belongs almost surely to an analytic tent space whenever $f\in \mathcal{AT}^p_{q,\alpha}$, and we also obtain the Triebel--Lizorkin counterpart. As part of the proof, we identify the random symbol space associated with $\mathcal{AT}^p_{q,\alpha}$ and solve the embedding problem from analytic tent spaces into mixed norm spaces. These results extend classical theorems of Hardy--Littlewood and Littlewood, as well as their later analogues for Bergman and mixed norm spaces.

math.FA

Rapid and robust parameter estimation for electrochemical battery models via BOLT: A batch-optimized local-to-global technique

Accurate and efficient parameter estimation is essential for applying electrochemical battery models in simulation, state estimation, control, and repeated model updating. However, conventional optimization methods, such as particle swarm optimization (PSO) and genetic algorithms (GA), often require many model evaluations and show considerable run-to-run variability, limiting their use in time-sensitive calibration scenarios. This study proposes a Batch-Optimized Local-to-Global Technique (BOLT) for rapid and robust parameter estimation of electrochemical battery models. BOLT combines diversified candidate initialization, batch-parallel trust-region reflective (TRF) local refinement, JIT-accelerated model evaluation, and multi-condition consistency screening within a unified calibration workflow. Comparative experiments based on a grouped single-particle model and measured data from a commercial 18650 NMC lithium-ion cell show that BOLT achieves a favorable trade-off among voltage-response accuracy, computational efficiency, and repeated-run stability. BOLT(32) achieves an average mean absolute error of \(12.4 \pm 0.1\) mV over five operating conditions, requiring only \(20636 \pm 3081\) model calls and \(8.97 \pm 1.20\) s per run. Synthetic-data validation with a known parameter vector in the grouped SPM formulation further shows that BOLT recovers the reference parameter vector under model-consistent conditions and remains robust under 1--3 mV voltage-noise perturbations, with the mean parameter absolute relative error below \(0.6\%\). These results indicate that BOLT provides a practical calibration framework for BMS parameter updating, control-oriented battery digital twins, and second-life battery screening.

eess.SY

Physics-guided residual Kalman learning for state-of-charge estimation of lithium iron phosphate batteries

Accurate state of charge (SOC) estimation of lithium iron phosphate (LFP) batteries remains challenging because of their flat open-circuit-voltage (OCV)-SOC characteristics, temperature-dependent dynamics, and sensitivity to initialization errors. Here, we propose a physics-guided residual Kalman learning (PRKL) framework for electrochemical-model-based SOC estimation. PRKL combines a control-oriented single-particle-model-based extended Kalman filter (EKF), which provides recursive physical state propagation, with a gated recurrent unit (GRU) residual learner that compensates structured EKF errors using electrochemical states and measurement features. The framework is evaluated on a public graphite/LFP dataset covering three dynamic drive cycles, eight temperatures from -10 to 50 degrees C, and initialization offsets up to 20 percent. Using dynamic stress test (DST) and federal urban driving schedule (FUDS) cycles for training and the supplemental federal test procedure (US06) cycle for cross-profile testing within the same cell dataset, PRKL achieves a global average root mean square error (RMSE) of 1.19 percent, corresponding to a 77 percent reduction relative to the physics-only EKF. These results show that electrochemical state information can guide residual learning and improve recursive SOC estimation for LFP batteries. The present validation supports cross-profile robustness within the studied dataset and provides a basis for future cross-cell, ageing-aware, and embedded-platform validation.

eess.SY

Two Problems in Bergman Spaces with Non-radial Weights

This paper investigates two problems unified by the study of the uniform boundedness of the dilation operators (UBD) T_r f(z)=f(rz), 0<r<1, acting on weighted Bergman spaces A^p_omega with not necessarily radial weights. We first characterize the random symbol space for A^p_omega under a mild admissible condition (Theorem 1.2). This extends the main result of [7] from radial weights to non-radial weights. We then introduce two new notions, namely non-radial mixed norm spaces M(p,q;omega) and analytic tent spaces A(p,q;omega), and we characterize their corresponding symbol spaces as well (Theorem 1.8, Theorem 1.12). The novelty here is to employ a measure-disintegration framework to prove a general Littlewood-type theorem (M(p,q;omega))* = H(2,q;omega_r), from which the Bergman space result (A^p_omega)* = H(2,p;omega_r) follows as the special case p=q. Among other things, UBD plays a pivotal role in the proofs of the preceding theorems. The second main problem addressed in this paper is to establish a local-to-global criterion for UBD, which remains largely unexplored for non-radial weights. Our principle result in this part (Theorem 1.16) asserts that UBD is guaranteed by two local geometric conditions: bounded hyperbolic oscillation (BHO) and a reverse-Carleson tail condition (RC). This is the technical heart of the paper. In the course of our investigation, three new types of problems arise naturally, each of independent interest: a two-weight top-maximal operator (Theorem 1.17); a truncated maximal operator over hyperbolic balls (Theorem 1.18); and a single-testing Carleson embedding problem (Theorem 1.19).

math.CV

RoSHAP: A Distributional Framework and Robust Metric for Stable Feature Attribution

Feature attribution analysis is critical for interpreting machine learning models and supporting reliable data-driven decisions. However, feature attribution measures often exhibit stochastic variation: different train--test splits, random seeds, or model-fitting procedures can produce substantially different attribution values and feature rankings. This paper proposes a framework for incorporating stochastic nature of feature attribution and a robust attribution metric, RoSHAP, for stable feature ranking based on the SHAP metric. The proposed framework models the distribution of feature attribution scores and estimates it through bootstrap resampling and kernel density estimation. We show that, under mild regularity conditions, the aggregated feature attribution score is asymptotically Gaussian, which greatly reduces the computational cost of distribution estimation. The RoSHAP summarizes the distribution of SHAP into a robust feature-ranking criterion that simultaneously rewards features that are active, strong, and stable. Through simulations and real-data experiments, the proposed framework and RoSHAP outperform standard single-run attribution measures in identifying signal features. In addition, models built using RoSHAP-selected features achieve predictive performance comparable to full-feature models while using substantially fewer predictors. The proposed RoSHAP approach improves the stability and interpretability of machine learning models, enabling reliable and consistent insights for analysis.

stat.ML

Seed Hijacking of LLM Sampling and Quantum Random Number Defense

Large language models (LLMs) rely on deterministic pseudorandom number generators (PRNGs) for autoregressive sampling, creating a critical supply-chain attack surface overlooked by existing defenses. We present SeedHijack, a backdoor attack that manipulates PRNG outputs to force attacker-specified token selection without altering model logits. In a 540-trial benchmark on GPT-2 (124M), the attack achieves 99.6% exact token injection across 9 sampling configurations; it reaches 100% success on four aligned models (1.5B-7B, RLHF/SFT/reasoning distillation) and bypasses all alignment methods tested in this work. We further propose a defense based on a hardware quantum random number generator (QRNG), which neutralizes the attack in our evaluated threat model with negligible median overhead (+0.6% latency, +7.7 MB memory). Our work identifies a critical sampling-layer vulnerability and provides a practical, deployable QRNG-based defense.

cs.CR

DPEPO: Diverse Parallel Exploration Policy Optimization for LLM-based Agents

Large language model (LLM) agents that follow the sequential "reason-then-act" paradigm have achieved superior performance in many complex tasks.However, these methods suffer from limited exploration and incomplete environmental understanding, as they interact with only a single environment per step. In this paper, we first introduce a novel paradigm that enables an agent to interact with multiple environments simultaneously and share cross-trajectory experiences. Building upon this paradigm, we further propose DPEPO, a reinforcement learning (RL) algorithm that encourages the agent to perform diverse parallel exploration. There are two stages in DPEPO: initial supervised fine-tuning (SFT) imparts basic parallel reasoning and action generation, followed by reinforcement learning stage with a hierarchical reward scheme. We design a parallel trajectory-level success reward and two step-level rewards: Diverse Action Reward and Diverse State Transition Reward, which actively penalize behavioral redundancy and promote broad exploration. Extensive experiments on ALFWorld and ScienceWorld show that DPEPO achieves state-of-the-art (SOTA) success rates, while maintaining comparable efficiency to strong sequential baselines. (Code is available at https://github.com/LePanda026/Code-for-DPEPO)

cs.CL

Turning Porous Functional Materials into Directional Transport Platforms with Unidirectional Surface Acoustic Waves

Porous media underpin absorption, filtration, separation, and high-area interfacial transport in chemical and diagnostic systems, yet sustained directional flow through them remains difficult because tortuous pore networks and strong acoustic losses promote bypassing, weak flow, and counterflow. Here, we show that floating-electrode unidirectional transducers (FEUDTs) convert porous materials into actively pumped transport platforms by generating predominantly unidirectional surface acoustic waves (SAWs) that couple more effectively than conventional interdigital transducers across wet multilayer interfaces. By varying pore size, permeability, sample thickness, and fluid viscosity, we find that transport is strongly enhanced when the SAW wavelength is comparable to the characteristic pore dimension, providing a practical design rule for acoustically activated porous media. Under these conditions, FEUDTs drive directional flow velocities up to 0.6 mm s$^{-1}$ at sub-watt input power, about 600 times faster than diffusion alone. FEUDTs also sustain pumping in prewetted porous media, where capillary contributions are removed, yielding velocities that exceed capillary-driven flow under matched conditions while remaining far above thermally induced transport. A reduced theoretical framework captures the main experimental trends and identifies transducer architecture, pore geometry, and actuation strength as the key parameters governing long-range, tunable transport in porous functional materials.

cond-mat.soft

FAVE: Flow-based Average Velocity Establishment for Sequential Recommendation

Generative recommendation has emerged as a transformative paradigm for capturing the dynamic evolution of user intents in sequential recommendation. While flow-based methods improve the efficiency of diffusion models, they remain hindered by the ``Noise-to-Data'' paradigm, which introduces two critical inefficiencies: prior mismatch, where generation starts from uninformative noise, forcing a lengthy recovery trajectory; and linear redundancy, where iterative solvers waste computation on modeling deterministic preference transitions. To address these limitations, we propose a Flow-based Average Velocity Establishment (Fave) framework for one-step generation recommendation that learns a direct trajectory from an informative prior to the target distribution. Fave is structured via a progressive two-stage training strategy. In Stage 1, we establish a stable preference space through dual-end semantic alignment, applying constraints at both the source (user history) and target (next item) to prevent representation collapse. In Stage 2, we directly resolve the efficiency bottlenecks by introducing a semantic anchor prior, which initializes the flow with a masked embedding from the user's interaction history, providing an informative starting point. Then we learn a global average velocity, consolidating the multi-step trajectory into a single displacement vector, and enforce trajectory straightness via a JVP-based consistency constraint to ensure one-step generation. Extensive experiments on three benchmarks demonstrate that Fave not only achieves state-of-the-art recommendation performance but also delivers an order-of-magnitude improvement in inference efficiency, making it practical for latency-sensitive scenarios.

cs.IR

Primal-dual dynamics featuring Hessian-driven damping and variable mass for convex optimization problems

This paper deals with a new Tikhonov regularized primal-dual dynamical system with variable mass and Hessian-driven damping for solving a convex optimization problem with linear equality constraints. The system features several time-dependent parameters: variable mass, slow viscous damping, extrapolation, and temporal scaling. By employing the Lyapunov analysis approach, we obtain the strong convergence of the trajectory generated by the proposed system to the minimal norm solution of the optimization problem, as well as convergence rate results for the primal-dual gap, the objective residual, and the feasibility violation. We also show that the convergence rates of the primal-dual gap, the objective residual, and the feasibility violation can be improved by appropriately adjusting these parameters. Further, we conduct numerical experiments to demonstrate the effectiveness of the theoretical results.

math.OC

MixerCSeg: An Efficient Mixer Architecture for Crack Segmentation via Decoupled Mamba Attention

Feature encoders play a key role in pixel-level crack segmentation by shaping the representation of fine textures and thin structures. Existing CNN-, Transformer-, and Mamba-based models each capture only part of the required spatial or structural information, leaving clear gaps in modeling complex crack patterns. To address this, we present MixerCSeg, a mixer architecture designed like a coordinated team of specialists, where CNN-like pathways focus on local textures, Transformer-style paths capture global dependencies, and Mamba-inspired flows model sequential context within a single encoder. At the core of MixerCSeg is the TransMixer, which explores Mamba's latent attention behavior while establishing dedicated pathways that naturally express both locality and global awareness. To further enhance structural fidelity, we introduce a spatial block processing strategy and a Direction-guided Edge Gated Convolution (DEGConv) that strengthens edge sensitivity under irregular crack geometries with minimal computational overhead. A Spatial Refinement Multi-Level Fusion (SRF) module is then employed to refine multi-scale details without increasing complexity. Extensive experiments on multiple crack segmentation benchmarks show that MixerCSeg achieves state-of-the-art performance with only 2.05 GFLOPs and 2.54 M parameters, demonstrating both efficiency and strong representational capability. The code is available at https://github.com/spiderforest/MixerCSeg.

cs.CV

MM-NeuroOnco: A Multimodal Benchmark and Instruction Dataset for MRI-Based Brain Tumor Diagnosis

Accurate brain tumor diagnosis requires models to not only detect lesions but also generate clinically interpretable reasoning grounded in imaging manifestations, yet existing public datasets remain limited in annotation richness and diagnostic semantics. To bridge this gap, we introduce MM-NeuroOnco, a large-scale multimodal benchmark and instruction-tuning dataset for brain tumor MRI understanding, consisting of 24,726 MRI slices from 20 data sources paired with approximately 200,000 semantically enriched multimodal instructions spanning diverse tumor subtypes and imaging modalities. To mitigate the scarcity and high cost of diagnostic semantic annotations, we develop a multi-model collaborative pipeline for automated medical information completion and quality control, enabling the generation of diagnosis-related semantics beyond mask-only annotations. Building upon this dataset, we further construct MM-NeuroOnco-Bench, a manually annotated evaluation benchmark with a rejection-aware setting to reduce biases inherent in closed-ended question formats. Evaluation across ten representative models shows that even the strongest baseline, Gemini 3 Flash, achieves only 41.88% accuracy on diagnosis-related questions, highlighting the substantial challenges of multimodal brain tumor diagnostic understanding. Leveraging MM-NeuroOnco, we further propose NeuroOnco-GPT, which achieves a 27% absolute accuracy improvement on diagnostic questions following fine-tuning. This result demonstrates the effectiveness of our dataset and benchmark in advancing clinically grounded multimodal diagnostic reasoning. Code and dataset are publicly available at: https://github.com/gfnnnb/MM-NeuroOnco

cs.CV

HalluGuard: Demystifying Data-Driven and Reasoning-Driven Hallucinations in LLMs

The reliability of Large Language Models (LLMs) in high-stakes domains such as healthcare, law, and scientific discovery is often compromised by hallucinations. These failures typically stem from two sources: data-driven hallucinations and reasoning-driven hallucinations. However, existing detection methods usually address only one source and rely on task-specific heuristics, limiting their generalization to complex scenarios. To overcome these limitations, we introduce the Hallucination Risk Bound, a unified theoretical framework that formally decomposes hallucination risk into data-driven and reasoning-driven components, linked respectively to training-time mismatches and inference-time instabilities. This provides a principled foundation for analyzing how hallucinations emerge and evolve. Building on this foundation, we introduce HalluGuard, an NTK-based score that leverages the induced geometry and captured representations of the NTK to jointly identify data-driven and reasoning-driven hallucinations. We evaluate HalluGuard on 10 diverse benchmarks, 11 competitive baselines, and 9 popular LLM backbones, consistently achieving state-of-the-art performance in detecting diverse forms of LLM hallucinations. We open-source our proposed \model{} model at https://github.com/Susan571/HalluGuard-ICLR2026.

cs.LG

From Lagging to Leading: Validating Hard Braking Events as High-Density Indicators of Segment Crash Risk

Identifying high crash risk road segments and accurately predicting crash incidence is fundamental to implementing effective safety countermeasures. While collision data inherently reflects risk, the infrequency and inconsistent reporting of crashes present a major challenge to robust risk prediction models. The proliferation of connected vehicle technology offers a promising avenue to leverage high-density safety metrics for enhanced crash forecasting. A Hard-Braking Event (HBE), interpreted as an evasive maneuver, functions as a potent proxy for elevated driving risk due to its demonstrable correlation with underlying crash causal factors. Crucially, HBE data is significantly more readily available across the entire road network than conventional collision records. This study systematically evaluated the correlation at individual road segment level between police-reported collisions and aggregated and anonymized HBEs identified via the Google Android Auto platform, utilizing datasets from California and Virginia. Empirical evidence revealed that HBEs occur at a rate magnitudes higher than traffic crashes. Employing the state-of-the-practice Negative-Binomial regression models, the analysis established a statistically significant positive correlation between the HBE rate and the crash rate: road segments exhibiting a higher frequency of HBEs were consistently associated with a greater incidence of crashes. This sophisticated model incorporated and controlled for various confounding factors, including road type, speed profile, proximity to ramps, and road segment slope. The HBEs derived from connected vehicle technology thus provide a scalable, high-density safety surrogate metric for network-wide traffic safety assessment, with the potential to optimize safer routing recommendations and inform the strategic deployment of active safety countermeasures.

cs.OH