arXiv subjects

Hongyu Liu

Publications and source records attributed to Hongyu Liu.

At least 37 records · Page 2Linked to original sources

arXiv2026

ThreshGuide: Class-Aware Labeled-Guided Thresholding for Semi-Supervised 3D Abdominal Multi-Organ Segmentation

Pseudo-labeling is a strong paradigm for semi-supervised medical image segmentation, yet its effectiveness is highly sensitive to confidence thresholding. In abdominal multi-organ segmentation, a fixed global threshold is particularly suboptimal because organ classes differ substantially in size, appearance, and learning difficulty. In this work, we propose ThreshGuide, a class-aware threshold adaptation framework that uses labeled data to guide pseudo-label selection on unlabeled data. Built upon a standard teacher-student architecture, the teacher model evaluates labeled samples during training to estimate class-aware threshold targets by maximizing an error-aware F\b{eta} criterion that balances precision and coverage. These targets are then smoothed with an exponential moving average (EMA) and used to filter unlabeled voxels in a class-dependent manner. Experiments on FLARE2022 and AMOS2022 show that ThreshGuide performs competitively overall, yielding clear improvements specifically on hard-to-learn organs.

cs.CV↗

arXiv2026

On passive recovery of structured elastic density and initial states

We study simultaneous recovery of the (variable) mass density, initial displacement, and initial velocity for the three-dimensional isotropic elastic wave equation with known constant Lamé parameters. The data are the complete displacement trace on an enclosing boundary. We first assume that the density-weighted initial displacement and velocity have fixed known profiles in one spatial direction. The $s^0$ and $s^1$ coefficients of the zero-frequency Laplace expansion identify these weighted states. The $s^2$ and $s^3$ coefficients then give static Lamé orthogonality identities for the density difference. The exact difference expansion through order $s^3$ has an $O(s^4)$ remainder, uniformly on bounded spatial sets and bounded density-contrast and weighted-state classes. Under alignment of the two initial states and a nonzero moment of the density-weighted initial velocity, the two density identities reduce to a constant-vector static transform. Two opposite elastic null phases give uniqueness for one or two fixed vertical density profiles when the profile family is two-sided Laplace nondegenerate. When $λ+μ\ne0$, each nonzero normal root has partial multiplicities $2$ and $1$ and admits a length-two Jordan chain. The associated polynomial--exponential Lamé mode produces derivatives of the bilateral profile transforms. The resulting Hermite--Laplace system gives uniqueness for aligned classes with up to four fixed vertical profiles and independent horizontal coefficients, provided that the profile system is Hermite--Laplace nondegenerate. Distinct translations of one compactly supported profile provide an explicit four-profile class. We also prove rigidity of the alignment reduction and exhibit an infinite-dimensional kernel for the reduced transform on unrestricted densities.

math.AP↗

arXiv2026

Boundary Spectral Inequalities from Measurable Sets on $W^{2,\infty}$ manifolds and their applications

Let $(\mathcal{M},g)$ be a compact connected $W^{2,\infty}$ Riemannian manifold with nonempty boundary and a uniformly elliptic Lipschitz metric. We establish quantitative boundary spectral inequalities for both Dirichlet and Neumann eigenfunctions from arbitrary measurable subsets of $\partial\mathcal{M}\times(-s_0,s_0)$ of positive surface measure. Moreover, we apply the Dirichlet spectral inequality to establish the boundary observability estimate \[ \|u(T)\|_{L^2(\mathcal{M})}^2 \le C \int_J \|\partial_{ν_g}u(x,t)\|^2 \,\mathrm{d}S_g\,\mathrm{d}t . \] where $J\subset\partial\mathcal{M}\times(0,T)$ is measurable and has positive surface--time measure. A localized theorem requiring $W^{2,\infty}$ regularity only near the observed boundary patch is retained as a separate consequence. The proof of the observability inequality combines quantitative continuation from positive-measure boundary sets with low-frequency spectral concentration. This result shows that boundary observability of solutions to the heat equation can also be achieved in $W^{2,\infty}$ domains. Finally, as an application, we prove that measurements of the boundary normal derivative on $J$ uniquely determine the initial state of heat equation. Under an a priori bound in $D(-Δ_{g, D}^{s/2})$, the recovery satisfies a logarithmic stability estimate of optimal order $s/2$.

math.AP↗

arXiv2026

Lipschitz Stability Estimates for Master Fields with Measure-Variable Regularity in Mean Field Games

In this paper, we investigate the Lipschitz stability of the master field and its functional derivative with respect to the measure variable for the master equation in mean field games. A key novelty of this work is the derivation of a nonlocal parabolic partial differential equation governing the functional measure derivative along a characteristic measure flow. Based on this equation, we establish a Carleman estimate for the difference of the measure derivatives associated with two sufficiently regular classical solutions. We also establish a Carleman estimate for the difference of the corresponding master fields, yielding a terminal-to-interior Lipschitz stability estimate along a characteristic measure flow. Together with stability of the associated measure flows, these estimates are used to control the resulting source terms. We then obtain a terminal-to-interior Lipschitz stability estimate for both the master field and its functional measure derivative in terms of the discrepancies of the terminal data. The result provides quantitative control of the master field with measure-variable regularity, and develops a Carleman-based framework for the stability analysis of master fields and their measure derivatives in mean field game master equations.

math.AP↗

arXiv2026

Parameter Identification in Reaction-Diffusion-Chemotaxis Systems from Single Turing Pattern Amplitudes

Turing patterns encode key information about biological mechanisms, yet traditional inverse problems rely on non-biological data like boundary measurements, neglecting the patterns themselves. We introduce a new direction that directly uses the amplitudes of Turing patterns for parameter identification. Motivated by stripe-like biological patterns, we study two 1D models: one with density-dependent chemotaxis ($χ(n,c)=χ_0 n$) and one with ratio-dependent chemotaxis ($χ(n,c)=χ_0 n/c$). We present a framework that uses the spatial amplitude profile of a stationary pattern to recover system parameters, including wavelength, diffusion constants, and chemotactic and kinetic coefficients---offering a biologically grounded paradigm for reverse-engineering pattern formation. The core contribution is a proof-of-concept showing that, under a Fourier truncation at mode $M=3$, the amplitude data can reconstruct the parameter combinations $\{d_n k^2,\, d_c k^2,\, χ_0 k^2,\, r\}$ for both models. These uniquely determine the ratios $d_c/d_n$ and $χ_0/d_n$, while individual parameters are recovered up to an overall scale. This is a purely theoretical work; numerical validation and stability analysis are left for future research. The framework is extensible to other models with suitable modifications.

math.AP↗

arXiv2026

Identifying unknown time delay and spatially varying coefficients in a reaction-diffusion equation from boundary measurements

This work investigates an inverse problem for a general class of linear reaction-diffusion systems incorporating multiple delayed contributions, namely retarded diffusion, retarded time derivatives, and retarded source terms. The objective is to simultaneously recover the unknown time lag $τ>0$ and spatially heterogeneous coefficients from boundary flux measurements alone. The recovery strategy exploits the singular temporal behavior generated by an incompatibility between the prescribed initial history and the boundary data. In contrast to prior inverse problems for delay equations, which assume $τ$ known and are confined to ODE or abstract settings, our approach operates in a parabolic PDE framework with spatially varying coefficients. Once $τ$ is identified, we establish a Lipschitz stability estimate via a Carleman inequality, in the case of time-independent coefficients $p=p(x)$, $q=q(x)$. This is the first result to simultaneously recover an unknown delay and spatially dependent coefficients in a delayed parabolic PDE from boundary data.

math.AP↗

arXiv2026

Superconducting Hydride Mg2RhH6 Experimentally Achieved at Lower Pressure

Although tremendous progress has been made in recent years in the field of polyhydride superconductors, the realization of high critical temperature superconductivity still relies on formidable high pressures. Searching for superconducting hydrides at lower pressures is of particular importance. Here we report the first experimental synthesis of the Mg2RhH6, which achieves superconductivity under a significantly reduced pressure of 30 GPa. The synthesis of Mg2RhH6 proceeds via a two step process (1) preparation of the Mg2RhH5 precursor containing hydrogen atoms stabilized by covalent bonds, followed by (2) hydrogen supplementation resulting in the filling of electrons into anti bonding orbitals above 30 GPa, which was accompanied by the structural transition from RhH5 square pyramid to RhH6 octahedron. Superconductivity is achieved at 30 GPa with a Tc of 24 K, which is further enhanced to 29 K at 53 GPa, evidenced by a sharp drop of resistivity to zero and characteristic suppression of Tc under applied magnetic fields. Our experiments prove the Mg2RhH6 superconductor to be thermodynamically stable above 30 GPa, making it the first case exhibiting a Tc of approximately 30 K at a readily accessible pressure. This study pioneers a highly promising pathway for the rational design and discovery of high temperature superconductors within phonon mediated BCS framework.

cond-mat.supr-con↗

arXiv2026

Exact high-order GPT cancellation in noncircular multilayers

This paper is concerned with exact high-order generalized polarization tensor (GPT) cancellation in two-dimensional noncircular multilayers with finite positive isotropic coating conductivities. We first prove that, for every admissible cyclically symmetric core--shell geometry, there exists a unique positive coating conductivity for which the structure is weakly neutral to every uniform incident field. For higher orders, we establish locally unique coating conductivities that cancel the complete $K$-contracted GPTs (K-CGPTs) block on a broad class of fixed noncircular multilayers with cyclic symmetry. This includes homothetic multilayers generated by strictly star-shaped symmetric domains, without assuming that the geometry is close to concentric disks. We also replace the near-background condition by a small-core condition covering finite, insulating, and perfectly conducting cores. By varying selected interface modes together with the coating conductivities near a radial GPT-vanishing structure, we further construct exact families with neither rotational nor reflection symmetry. For three coatings, the relevant nondegeneracy conditions and an explicit chiral example are rigorously certified. Finally, we show that the resulting cancellation yields enhanced near-cloaking for every entire harmonic incident field. A set of certified numerical examples fully corroborates the theoretical predictions established in this work.

math.AP↗

arXiv2026

Analytical Resource Management for Fine-grained MoE Computation-Communication Overlap

Fine-grained computation--communication overlap in distributed Mixture-of-Experts (MoE) inference allows communication to begin as partial compute results become ready. However, cooperative thread arrays (CTAs) performing computation and communication contend for finite residency capacity on streaming multiprocessors (SMs). Because a resident CTA generally retains its allocated SM resources until completion, CTAs that cannot be co-resident must wait for resources, resulting in wave-like execution. A fixed resource partition cannot adapt to changes in input size, routed expert load, and kernel configuration, potentially causing a communication backlog or reducing expert compute parallelism. We present a wave-quantized analytical model and launch-time resource manager for dependency-coupled overlap pipelines. Using routed-tile counts, kernel occupancy, GPU residency constraints, and split-level readiness dependencies, it selects the communication-CTA count and resource partition before each launch without candidate execution, per-workload profiling, or kernel recompilation. We integrate the method into the public COMET A100 implementation in FLUX. We evaluate three MoE models on four NVIDIA A100 GPUs under several parallelism strategies at the GEMM2+GatherRS operator, complete post-router MoE layer, and complete-model prefill levels. Across 15 real-p90 workloads, the analytical selector achieves 3.22 percent mean regret relative to the measured oracle with a mean solver overhead of 0.157 microseconds. Over COMET, our method achieves geometric-mean speedups of 2.528x at the GEMM2+GatherRS operator, 1.771x at the complete post-router MoE layer, and 1.185x for complete-model prefill, with maxima of 4.218x, 2.584x, and 1.439x, respectively. At every feasible TP=2/EP=2 sequence length of at least 4,096, our implementation outperforms COMET, Megatron core-TE, and FastMoE TP+NCCL.

cs.DC↗

arXiv2026

VoxPrivacy: A Benchmark for Evaluating Interactional Privacy of Speech Language Models

As Speech Language Models (SLMs) transition from personal devices to shared, multi-user environments such as smart homes, a new challenge emerges: the model is expected to distinguish between users to manage information flow appropriately. Without this capability, an SLM could reveal one user's confidential schedule to another, a privacy failure we term interactional privacy. Thus, the ability to generate speaker-aware responses becomes essential for SLM safe deployment. Current SLM benchmarks test dialogue ability but overlook speaker identity. Multi-speaker benchmarks check who said what without assessing whether SLMs adapt their responses. Privacy benchmarks focus on globally sensitive data (e.g., bank passwords) while neglecting contextual privacy-sensitive information (e.g., a user's private appointment). To address this gap, we introduce VoxPrivacy, the first benchmark designed to evaluate interactional privacy in SLMs. VoxPrivacy spans three tiers of increasing difficulty, from following direct secrecy commands to proactively protecting privacy. Our evaluation of nine SLMs on a 32-hour bilingual dataset reveals a widespread vulnerability: most open-source models perform close to random chance (around 50% accuracy) on conditional privacy decisions, while even strong closed-source systems fall short on proactive privacy inference. We further validate these findings on Real-VoxPrivacy, a human-recorded subset, confirming that failures observed on synthetic data persist in real speech. Finally, we demonstrate a viable path forward: by fine-tuning on a new 4,000-hour training set, we improve privacy-preserving abilities while maintaining robustness. To support future work, we release the VoxPrivacy benchmark, the large-scale training set, and the fine-tuned model to foster the development of safer and more context-aware SLMs.

eess.AS↗

arXiv2026

Superconductivity of Tellurium Polyhydride with Tc above 90K

We report experimental diacovery of superconductivity (SC) in tellurium (Te) polyhydride. The compound was synthesized at high pressure and high temperature conditions using a diamond anvil cell combined with a laser heating system. Subsequent in situ transport measurements at high pressures, performed as a function of temperature and applied magnetic field, revealed a superconducting transition with a critical temperature Tc about 91 K at 263 GPa. The superconducting phase is assigned to TeH4 with characterized face shared TeH12 cage forming quasi molecular H2 units based on synchrotron x-ray diffraction experiments. Analysis of the SC behavior at magnetic fields yielded a Ginzburg Landau (GL) coherence length of approximately 54 angstroms. Tellurium polyhydride thus becomes another chalcogen polyhydride superconductor in addition to the landmark discovery of the first polyhydride high Tc SC SH3.

cond-mat.supr-con↗

arXiv2026

Time-modulated refocusing in inhomogeneous medium

In an inhomogeneous dissipative medium, different rays experience different round-trip attenuation, so an instantaneous time mirror may retrace rays without producing a balanced point focus. We solve this problem for a microlocally selected back-propagating transverse-magnetic component generated by a short pulsating modulation of a thin shell surrounding an inhomogeneous Debye medium with Ohmic conductivity. An exact decomposition of the Maxwell--Debye field isolates the pulse-generated return, and a semiclassical mode reduction yields a scalar returned operator whose principal symbol separates the temporal pulse response, shell modulation, and accumulated material loss. Our analysis shows that geometric distortion and amplitude attenuation are both involved in wave propagation, but the geometric distortion cancels between the outgoing and returning branches, whereas physical attenuation remains. A normalized directional energy, computed from pre-modulation measurements using the known lossless backward propagator, contains the same leading attenuation factor as the returned symbol. Its reciprocal defines a smooth, nonnegative, physically admissible modulation that equalizes the selected source-side gain up to a relative $O(h)$ error without requiring a complete attenuation model. With such a modulation, the field generated by a point source has a focal-scale leading profile independent of directional attenuation, with a unique maximum at the source and coherent return time. In every fixed focal window, the maximum is displaced by at most $O(h^2)$ in space and time, and the spatial half-amplitude area is $O(h^2)$.

math.AP↗

arXiv2026

Global recovery of Lorentzian principal geometry from the hyperbolic Dirichlet-to-Neumann map

We prove that the full Dirichlet-to-Neumann map on a finite time interval globally determines the Lorentzian metric $\mathbf G_{g,c}=-c(x,t)^2\mathrm{d}t^2+g(x)$ encoded by the principal symbol of the wave operator. The result holds in spatial dimensions $n\geq3$ for multiplicatively separable wave speeds $c(x,t)=a(x)b(t)$, provided the accumulated effective time during the experiment exceeds the maximal travel time from the boundary to the interior and back. The data determine both the unknown Riemannian metric $g(x)$ and the wave speed $c(x,t)$ throughout the observation cylinder, up to a spatial change of coordinates that fixes the boundary and leaves the measured time unchanged. In particular, the temporal factor $b$ is not prescribed but is recovered from the same measurements. We also construct several examples demonstrating that the dimensional and visibility conditions for uniqueness are sharp.

math.AP↗

arXiv2026

Detection of a moving point-like scatter by using moving receivers and emitter

We consider an inverse scattering problem for the scalar wave equation in which the point emitter, a point-like scatterer, and multiple receivers are all moving. The goal is to reconstruct the trajectory of a moving point-like scatterer from time-dependent measurements of the scattered field. To this end, we establish a rigorous point-interaction model for the moving scatterer in the time domain, proving well-posedness of the forward scattering problem, including existence, uniqueness and continuous dependence of the scattered field on the scatterer trajectory and scattering parameter. By exploiting the retarded-time structure, we introduce distance functions that connect emission, scattering, and observation processes, and show that they satisfy a coupled system of nonlinear ordinary differential equations determined by the measurement data. Based on this formulation, we propose a reconstruction algorithm that combines initial localization with the solution of the derived ODE system. Numerical experiments demonstrate that the method is accurate and robust with respect to noise, with reconstruction errors remaining stable under moderate perturbations.

math.AP↗

arXiv2026

Determining internal topological structures and running cost of mean field games with partial boundary measurement

This paper investigates the simultaneous reconstruction of the running cost function and the internal topological structure within the mean-field games (MFG) system utilizing partial boundary data. The inverse problem is notably challenging due to factors such as nonlinear coupling, the necessity for multi-parameter reconstruction, constraints on probability measures, and the limited availability of measurement information. To address these challenges, we propose an innovative approach grounded in a higher-order linearization method. This method is tailored for inverse problems in MFG systems that involve Dirichlet and Neumann boundary conditions. Initially, we present unique reconstruction results for the cost function and internal topological structure of the MFG system under various homogeneous boundary conditions. Subsequently, we extend these results to accommodate inhomogeneous boundary conditions. These findings greatly enhance our understanding of simultaneous reconstruction in complex MFG systems.

math.OC↗

arXiv2026

Derivative-Free Recovery of a Nonlinearity in a Free-Boundary DCIS Model

We investigate an inverse coefficient problem for a multidimensional free-boundary model of ductal carcinoma in situ (DCIS), in which the tumor interface is governed by the nonlinear coupling of nutrient concentration, tissue pressure and curvature, and the unknown nutrient consumption function is recovered from a temporal trace of the nutrient concentration obtained by needle aspiration biopsy. For the forward problem, we establish uniform local well-posedness over an admissible class of consumption functions. The inverse problem is recast as a fixed-point problem: approximating the admissible set by finite-dimensional spaces yields discrete iteration operators, for which we prove the existence of fixed points, and the strong convergence of a subsequence of discrete fixed points to a fixed point of the continuous operator, which solves the inverse problem under a consistency condition. To approximate these fixed points, we develop a homotopy-continuation method combining a linearly convergent Picard iteration with a cubical Sperner search, without differentiating an objective functional or computing an adjoint state. Several numerical experiments on radially symmetric and non-symmetric DCIS models corroborate the theoretical findings.

math.NA↗

arXiv2026

A Weyl-type law for surface-localized eigenfunctions of the Maxwell transmission problem

This work studies surface-localized transmission eigenfunctions for time-harmonic electromagnetic scattering. Prior results on this phenomenon are mostly qualitative, proving existence without quantifying distribution. Here we provide a quantitative analysis for the Maxwell transmission eigenvalue problem, covering both TE and TM modes. We first establish a Weyl asymptotic law for the full counting function, with cubic growth with respect to the radius $R$. We then prove Weyl-type upper and lower bounds for the counting functions restricted to surface-localized modes, and demonstrate that they share the same growth order three. To our knowledge, this is the first quantitative result of its kind, demonstrating that surface-localized eigenfunctions form a non-negligible fraction of the high-frequency spectrum.

math.AP↗

arXiv2026

Inverse Scattering from Conformal Infinity for Totally Geodesic Defects in Hyperbolic Space

We study inverse scattering from conformal infinity for impenetrable topological defects whose interaction surfaces are totally geodesic in hyperbolic space $\mathbb H^n$, $n \geq 2$. For a fixed spectral parameter $λ_0 > 0$, the prescribed inputs are boundary labels $ξ\in \partial_\infty \mathbb H^n$. Each label selects an incoming Helgason mode in the hyperbolic interior. Given a defect $\mathcal P \Subset \mathbb H^n$, this mode generally fails to satisfy the homogeneous trace condition on the interaction surface and hence generates an outgoing correction. The leading coefficient of this correction at conformal infinity defines the measured far-field pattern. The central question is whether $\mathcal P$ can be recovered from the far-field patterns corresponding to one or finitely many prescribed boundary labels. This gives a formally determined inverse problem at one fixed spectral parameter. Our first main result establishes the unique determination of totally geodesic defects, an admissible class that includes both bulk components and hypersurface-supported ones. For Dirichlet-type defects, a single boundary label suffices. For Neumann-type defects, $n+1$ boundary labels are sufficient and in general necessary. These labels are required to satisfy the natural affine-independence condition at conformal infinity. Our second main result provides quantitative stability estimates within the same framework. The hyperbolic Hausdorff distance between two defects is controlled by the discrepancy of their far-field patterns at conformal infinity. The proof combines continuation from conformal infinity with quantitative geodesic reflection across totally geodesic hypersurfaces. Taken together, these results yield a qualitative and quantitative far-field inverse scattering theory at conformal infinity, based on formally determined data.

math.AP↗