arXiv Science⌕ Search

arXiv subjects

Ting Zhou

Publications and source records attributed to Ting Zhou.

At least 19 recordsLinked to original sources

Simultaneous recovery of elastic and viscous coefficients in a Kelvin-Voigt equation

In this paper, we study the simultaneous recovery of a spatially dependent elastic coefficient and a time- and space-dependent viscosity coefficient in a scalar Kelvin-Voigt equation. We prove that the displacement-to-traction map over any fixed time interval uniquely determines both coefficients, provided that the Riemannian metric induced by the viscosity is simple at each time. In the velocity formulation, the elastic term gives rise to a Volterra memory term involving second-order spatial derivatives, with a corresponding memory contribution to the boundary traction. We first determine all normal derivatives of both coefficients at the boundary using highly oscillatory boundary data. We then construct exact parabolic exponential solutions in the presence of memory and determine the viscosity by boundary rigidity. With the viscosity determined, an approximate identity argument reduces recovery of the elastic coefficient to injectivity of the scalar geodesic ray transform. Finally, numerical experiments illustrate a regularized finite-dimensional reconstruction of both coefficients from finitely many moments of the displacement-to-traction map.

math.AP↗

Spectrality of convolution and product measures: a block decomposition approach

We study the spectrality of convolution and product measures. Let~$μ$~be a compactly supported Borel probability measure on~$\Bbb{R}^d$~such that ~$\mathcal{Z}(\widehatμ)\cup\{0\}=T$~is an additive subgroup, and let~$ν$~be a Borel probability measure on~$\Bbb{R}^d$. We prove that~$μ*ν$~is spectral if and only if~$T$~is a spectrum of~$μ$, ~$ν$~is spectral, and~$suppν$~is contained in a translate of the annihilator of~$T$. Moreover, every spectrum of~$μ*ν$~decomposes into full cosets of~$T$, and every choice of one point from each coset is a spectrum of~$ν$. We also obtain a convolution-factor criterion which strengthens a result of Wu. As applications, we characterize the spectrality of~$L_{[0,1]}*ν$~and prove that the spectrality of~$L_{[0,1]^d}*ν$~implies that of~$ν$~without a compact support assumption. For product measures, we give necessary and sufficient conditions for an orthogonal set to be a spectrum by relating its block structure to spectra of the factors. No compact support assumptions are required.

math.FA↗

SCOPE-AD: Sequential cost-aware ordinal-belief planning with energy-based models for diagnostic agents

Alzheimer's disease (AD) diagnosis requires sequential evidence acquisition under heterogeneous test costs and patient burden. Fixed-modality predictors do not jointly decide which test to acquire or when the available evidence is sufficient for diagnosis. We propose SCOPE-AD (Sequential Cost-Aware Ordinal-Belief Planning with Energy-Based Models for Diagnostic Agents) for cost-aware classification of cognitively normal (CN), mild cognitive impairment (MCI), and AD cases. A mask-aware ordinal model represents uncertainty along the ordered CN--MCI--AD continuum. Retrospective training records provide sampled Bellman targets for an energy-based teacher, whose action distributions are distilled into a Qwen policy. At deployment, the agent selects acquisition or diagnosis actions under availability and budget constraints without access to unacquired values. After each acquisition, the evidence and ordinal belief are updated before the next decision. On ADNI, SCOPE-AD achieves 77.70\% Macro-F1 at an average acquisition cost of \$50.46, exceeding the strongest evaluated baseline by 9.34 percentage points. Full-modality evaluation raises Macro-F1 by only 1.89 points while increasing acquisition cost by 116.7 times. These results support selective acquisition for cost-effective diagnosis.

cs.AI↗

Stable determination of coefficients for the nonlinear third-order acoustic equations

In this paper, we study an inverse boundary value problem for a generalized Jordan-Moore-Gibson-Thompson equation with Westervelt-type nonlinearity and space-dependent coefficients. This third-order (in time) hyperbolic equation models nonlinear ultrasound propagation in viscous and thermally relaxing media. Assuming that the diffusivity and the sound speed are known a priori, we investigate the simultaneous stable determination of the friction coefficient, the weak damping coefficient, the potential, and the nonlinear coefficient from the associated Dirichlet-to-Neumann map. The proof combines the finite-difference linearization method with the construction of Gaussian beam and geometric optics solutions. These arguments reduce the inverse problem to stability estimates for (attenuated) geodesic ray transforms under the foliation condition. As a consequence, we obtain Hölder-type stability estimates of recovering the linear and nonlinear coefficients.

math.AP↗

Logarithmic Stability for an Inverse Source Problem in a Coupled Nonlinear Helmholtz System

We study an inverse source problem for a two-mode nonlinear Helmholtz system motivated by second-harmonic generation. The datum is the full first Fréchet derivative of the nonlinear Dirichlet-to-Neumann map at one fixed, common, small boundary state, which need not be zero. For real-valued small data, a known susceptibility bounded away from zero, and sources supported in a fixed compact subset of the domain, we prove uniqueness and a conditional single-logarithmic stability estimate. The first boundary variation is the Dirichlet-to-Neumann map of a symmetric $2\times2$ matrix Schrödinger operator. Absorbing the two known Helmholtz energies into its matrix potential permits the use of standard zero-energy complex geometrical optics solutions with a common null phase geometry. A bilinear Alessandrini identity then gives a uniform Fourier estimate for the two entries containing the background fields. Zero extension below the exponent $3/2$, a low/high frequency splitting, and interpolation between $H^{-2}$ and the a priori $H^s$ source bound yield the stated stability modulus.

math.AP↗

An Inverse Random Source Problem for the Moore-Gibson-Thompson Equation Driven by Fractional Brownian Motion

In this paper, we consider an inverse random source problem for the stochastic Moore-Gibson-Thompson equation driven by fractional Brownian motion with Hurst index $H \in (0, 1)$ of the form $f_1(x)g_1(t)\dot{B}^H(t)+f_2(x)g_2(t)$. Given the random source, existence and uniqueness of mild solutions are verified. For the inverse problem, the uniqueness of recovering the strength $f_i(x)$ if the time functions $g_i$ are known and $g_i(t)$ if the spatial functions $f_i$ are known when $H \in(0,1)$ from the boundary flux on a special nonempty open subset is proved.

math.AP↗

Beyond Solvability: Task Learnability as a Static Prior for LLM RL Post-Training

Reinforcement learning (RL) has become a central post-training paradigm for eliciting reasoning capabilities in large language models, yet uniform task sampling allocates compute without regard to differences in how tasks respond to optimization. Existing task-valuation methods mostly rely on snapshot-based signals such as current pass rate or reward, which estimate how solvable a task is under the current policy. However, tasks with similar current solvability can still differ substantially in how positively they respond to further training. We study this residual axis as task learnability: a regime-conditional measure of expected positive response to continued training under a fixed RL post-training regime. By analyzing per-task reward trajectories, we find that learnability is reproducible across independently sampled training contexts and predictive of downstream utility. To make this signal practical before training begins, we propose TrajVal, a lightweight probe-based estimator that approximates per-task learnability from a short probe run and two endpoint evaluations. TrajVal can be used either as a standalone static prior for task sampling or as a multiplicative prior for existing online schedulers. Experiments on mathematical and logical reasoning benchmarks across multiple model scales show that TrajVal improves data efficiency over uniform sampling and provides complementary gains when combined with online scheduling methods.

cs.LG↗

Partial data Calderón problem for quasilinear conductivities in dimension 2

In this paper, we prove a uniqueness result for the partial data Calderón problem with quasilinear conductivity in two dimensions. The proof is based on higher-order linearization and the use of CGO solutions in dimension two that vanish on part of the boundary. Since derivatives of the solutions appear in the integral identity, we need improved remainder estimates, for which we introduce a modification in the choice of the phase, analogous to limiting Carleman weights. We also analyze how combinations of phases produce specific patterns in the products of solutions, which allows us to apply both stationary and nonstationary phase arguments to recover the conductivity.

math.AP↗

Semantic Noise Aided Secure Image Transmission over MIMO Fading Channels

Existing semantic communications have exhibited satisfactory performance in many tasks, but secure image transmission remains insufficiently explored. We propose a novel secure image semantic communication (SISC) framework over multiple-input multiple-output (MIMO) fading channels. To ensure high-quality image reconstruction for the legitimate semantic user (SU) and simultaneously interfere with the eavesdropper (Eve), we design a semantic noise generation (SNG) network. This network generates a beneficial semantic noise map based on both the source features and the SU channel state information (CSI). An efficient channel estimation enhanced network is incorporated to obtain the accurate CSI and enhance the system performance. Furthermore, to improve the secure image reconstruction quality, we develop an efficient transceiver beamformer optimization algorithm, where the formulated problem is solved using the constrained stochastic successive convex approximation method. In the proposed SISC framework, semantic noise generation and beamforming optimization work together to ensure secure and high-quality image transmission. Numerical results demonstrate that the proposed semantic noise aided transmission scheme effectively protects image information from leakage to Eve while maintaining high-fidelity image reconstruction at SU.

cs.IT↗

GEOALIGN: Geometric Rollout Curation for Robust LLM Reinforcement Learning

Online reinforcement learning is widely used to align large language models (LLMs) with reward signals, yet training can be unstable under noisy or misspecified rewards. We identify a failure mode we call directional inconsistency: within a batch, a small set of high-reward rollouts induces representation-space preference directions that sharply disagree with the batch majority, resulting in high-variance and destabilizing updates. We propose geoalign, a lightweight plug-in for rollout curation in iterative policy optimization. Geoalign (i) forms within-prompt preference pairs, (ii) learns an online projector on per-rollout hidden states to concentrate reward-ordered displacement directions, and (iii) detects directionally inconsistent rollouts via their angular deviation from a batch consensus prototype and rectifies them with within-prompt stable alternatives. Geoalign is forward-pass only and adds negligible overhead. Across dialogue alignment with a learned reward model and mathematical reasoning with binary verified rewards, Geoalign improves final performance and reduces training oscillation, outperforming PF-PPO, PAR, PODS, and Seed-GRPO. These results suggest latent directional consensus as an effective reliability signal for online LLM RL.

cs.LG↗

Gauge symmetry and uniqueness in inverse problems for the JMGT equation

In this paper, we study an inverse boundary value problem for the Jordan--Moore--Gibson--Thompson equation on a simple Riemannian manifold. We consider an all boundary measurement map that maps Dirichlet boundary data and initial data to the corresponding Neumann-type boundary data and final-time data. Our main result shows that the nonlinear acoustic coefficient $β$ is uniquely determined by this measurement map, and the linear damping coefficients $α$ and $q$, along with the internal source term $F$, can be recovered up to a gauge symmetry. As a corollary, we also establish a specific case in which all coefficients are uniquely recovered. The proof relies on the method of first-order and second-order linearization and on the construction of geometric optics solutions. In the intermediate step, we establish the unique recovery of the lower-order coefficients in the linearized MGT equation.

math.AP↗

HumanVBench: Probing Human-Centric Video Understanding in MLLMs with Automatically Synthesized Benchmarks

Evaluating the nuanced human-centric video understanding capabilities of Multimodal Large Language Models (MLLMs) remains a great challenge, as existing benchmarks often overlook the intricacies of emotion, behavior, and cross-modal alignment. We introduce HumanVBench, a comprehensive video benchmark designed to rigorously probe these capabilities across 16 fine-grained tasks. A cornerstone of our work is a novel and scalable benchmark construction methodology, featuring two automated pipelines that synthesize high-quality video annotations and challenging multiple-choice questions with minimal human labor. By leveraging state-of-the-art models for annotation and systematically converting model-induced errors into plausible distractors, our framework provides a generalizable ``machine'' for creating nuanced evaluation suites. Our extensive evaluation of 30 leading MLLMs on HumanVBench reveals critical deficiencies, particularly in perceiving subtle emotions and aligning speech with visual cues, with even top proprietary models falling short of human performance. We open-source HumanVBench and our synthesis pipelines to catalyze the development of more socially intelligent and capable video MLLMs.

cs.CV↗

Inverse boundary value problems of determining nonlinear coefficients for the JMGT equation

We consider inverse boundary value problems for the Jordan-Moore-Gibson-Thompson (JMGT) equation in nonlinear acoustics with quadratic nonlinearities of Kuznetsov-type and Westervelt-type. We show that the associated boundary Dirichlet-to-Neumann map uniquely determines the nonlinear coefficients $β$ in the Westervelt-type model, and the pair $(β,κ)$ in the Kuznetsov-type model, provided that the observation time is greater than the maximal boundary-to-boundary geodesic travel time. The results are obtained in both the Euclidean setting and on compact Riemannian manifolds with proper geometric assumptions. The proof is based on the idea of second order linearization combined with the construction of geometric optics and Gaussian beam solutions, reducing the inverse problem of uniqueness to the injectivity of associated geodesic ray transforms.

math.AP↗

Uniqueness Result For Semi-linear Wave Equations With Sources

This paper addresses the inverse problem of simultaneously recovering multiple unknown parameters for semilinear wave equations from boundary measurements. We consider an initial-boundary value problem for a wave equation with a general semilinear term and an internal source. The inverse problem is to determine the nonlinear coefficients (potentials), the source term, and the initial data from the Dirichlet-to-Neumann (DtN) map. Our approach combines higher-order linearization and the construction of complex geometrical optics (CGO) solutions. The main results establish that while unique recovery is not always possible, we can precisely characterize the gauge equivalence classes in the solutions to this inverse problem. For a wave equation with a polynomial nonlinearity of degree $n$, we prove that only the highest-order coefficient can be uniquely determined from the DtN map; the lower-order coefficients and the source can only be recovered up to a specific gauge transformation involving a function $ψ$. Furthermore, we provide sufficient conditions under which unique determination of all parameters is guaranteed. We also extend these results to various specific non-polynomial nonlinearities, demonstrating that the nature of the nonlinearity critically influences whether unique recovery or a gauge symmetry is obtained.

math.AP↗

Exploring Constraints on Axion-like Particles with the Observations for blazar Mrk 501

Oscillations between axion-like particles (ALPs) and photons in astrophysical magnetic fields can lead to irregularities in the high energy gamma ray spectra of blazars. The magnetic field within the blazar jet plays a crucial role in shaping these effects, with its strength in the emission region being an important parameter determined by multi-wavelength observations. However, the origin of the high energy bump observed in the spectral energy distribution of some blazars is a topic of debate, with both leptonic and hadronic scenarios providing plausible explanations that result in different magnetic field strengths in the emission region. In this study, we investigate the impact of magnetic field configurations on the constraints of ALP parameters. We consider both leptonic and hadronic emission scenarios for the blazar Mrk 501 and derive the corresponding jet magnetic field strengths. Additionally, we explore two jet magnetic field models: one with a toroidal component and the other with helical and tangled components. By analyzing the spectra of Mrk 501 observed by MAGIC and Fermi-LAT from 2017-06-17 to 2019-07-23, which are well-described by both emission scenarios, we derive constraints on the ALP parameters. Our results demonstrate that both the emission scenario and the magnetic field structure play a significant role in deriving these constraints, with the hadronic model leading to much more stringent limits compared to the leptonic model.

astro-ph.HE↗

Bridging BCI and Communications: A MIMO Framework for EEG-to-ECoG Wireless Channel Modeling

As a method to connect human brain and external devices, Brain-computer interfaces (BCIs) are receiving extensive research attention. Recently, the integration of communication theory with BCI has emerged as a popular trend, offering potential to enhance system performance and shape next-generation communications. A key challenge in this field is modeling the brain wireless communication channel between intracranial electrocorticography (ECoG) emitting neurons and extracranial electroencephalography (EEG) receiving electrodes. However, the complex physiology of brain challenges the application of traditional channel modeling methods, leaving relevant research in its infancy. To address this gap, we propose a frequency-division multiple-input multiple-output (MIMO) estimation framework leveraging simultaneous macaque EEG and ECoG recordings, while employing neurophysiology-informed regularization to suppress noise interference. This approach reveals profound similarities between neural signal propagation and multi-antenna communication systems. Experimental results show improved estimation accuracy over conventional methods while highlighting a trade-off between frequency resolution and temporal stability determined by signal duration. This work establish a conceptual bridge between neural interfacing and communication theory, accelerating synergistic developments in both fields.

eess.SP↗

Computational fluid dynamics-based structure optimization of ultra-high-pressure water-jet nozzle using approximation method

Since the geometry structure of ultra-high-pressure (UHP) water-jet nozzle is a critical factor to enhance its hydrodynamic performance, it is critical to obtain a suitable geometry for a UHP water jet nozzle. In this study, a CFD-based optimization loop for UHP nozzle structure has been developed by integrating an approximate model to optimize nozzle structure for increasing the radial peak wall shear stress. In order to improve the optimization accuracy of the sparrow search algorithm (SSA), an enhanced version called the Logistic-Tent chaotic sparrow search algorithm (LTC-SSA) is proposed. The LTC-SSA algorithm utilizes the Logistic-Tent Chaotic (LTC) map, which is designed by combining the Logistic and Tent maps. This new approach aims to overcome the shortcoming of "premature convergence" for the SSA algorithm by increasing the diversity of the sparrow population. In addition, to improve the prediction accuracy of peak wall shear stress, a data prediction method based on LTC-SSA-support vector machine (SVM) is proposed. Herein, LTC-SSA algorithm is used to train the penalty coefficient C and parameter gamma g of SVM model. In order to build LTC-SSA-SVM model, optimal Latin hypercube design (Opt LHD) is used to design the sampling nozzle structures, and the peak wall shear stress (objective function) of these nozzle structures are calculated by CFD method. For the purpose of this article, this optimization framework has been employed to optimize original nozzle structure. The results show that the optimization framework developed in this study can be used to optimize nozzle structure with significantly improved its hydrodynamic performance.

physics.flu-dyn↗

Comprehensive Optimization of Interferometric Diffusing Wave Spectroscopy (iDWS)

It has been shown that light speckle fluctuations provide a means for noninvasive measurements of cerebral blood flow index (CBFi). While conventional Diffuse Correlation Spectroscopy (DCS) provides marginal brain sensitivity for CBFi in adult humans, new techniques have recently emerged to improve diffuse light throughput and thus, brain sensitivity. Here we further optimize one such approach, interferometric diffusing wave spectroscopy (iDWS), with respect to number of independent channels, camera duty cycle and full well capacity, incident power, noise and artifact mitigation, and data processing. We build the system on a cart and define conditions for stable operation. We show pulsatile CBFi monitoring at 4-4.5 cm source-collector separation in adults with moderate pigmentation (Fitzpatrick 4). We also report preliminary clinical measurements in the Neuro Intensive Care Unit (Neuro ICU). These results push the boundaries of iDWS CBFi monitoring performance beyond previous reports.

physics.med-ph↗