Papers

Despite significant progress in visual tasks by Multimodal Large Language Models (MLLMs), geometric diagram understanding remains challenging due to the presence of sparse visual cues and ambiguous symbol-primitive associations. MLLMs may therefore rely on textual priors, producing interpretations that conflict with visual evidence. We introduce the training-free Criticality-Driven Visual Intervention Framework (CVIF), an inference-time method that localizes critical layers and executes visual interventions during the transition from evidence aggregation to semantic decoding. At these layers, a Geometry-Constrained Local Relation Reconstruction (GCLR) module selects and weights vertex-centered visual evidence, while an Adaptive Visual Steering Operator (AVSO) redistributes attention mass toward the selected tokens. Experiments on PGPS9K and PGDP5K show that CVIF raises Overall F1 from 77.85 to 85.58 and from 75.23 to 82.84, respectively, establishing a novel inference-time visual intervention paradigm.

Gliomas are the most common type of primary brain tumors. Diffuse low-grade gliomas (DLGGs) are slow-growing tumors that are often minimally symptomatic for a long period of time. They progress to a higher grade, resulting in the patient's death. Treatments include surgery, chemotherapy and radiation therapy to control tumor progression. Responses to treatments are highly variable among patients. Furthermore, the diffuse component of this tumor entity has been observed but is not yet well measured. Here, we propose a new variable to quantify glioma morphology, called ESVR (Extra Sphere Volume Ratio). We use then a machine learning approach to study the importance of different variables for discriminating patterns of patients: patient-specific variables, genetic alterations of tumor tissue, and some image-based variables like ESVR and the volume of the tumor, obtained by MRI analysis. Our machine learning approach shows that the patient's age and ESVR at diagnosis, as well as the pathology results seem to play an important role. Taking the highlighted variables into account could thus help to model the tumor behavior. In association with other classical biomarkers, our new ESVR feature could help the clinician to manage the treatment, as it carries significant information in terms of aggressivity of the DLGG.

We propose MOSEL (Multi-Objective Stackelberg Efficient Learning), a framework for a posteriori multi-objective optimization (MOO) in deep neural networks that recovers a full front of Pareto stationary solutions at the computational cost of standard single-objective training. MOSEL reformulates the problem as a bilevel optimization problem that leverages network modularity to decouple representation learning from objective-preference alignment. Casting the bilevel optimization problem as a Stackelberg game enables solving the original a posteriori MOO problem in a single forward-backward pass. As a result, MOSEL matches the time and memory efficiency of standard single-objective training while enabling scalable Pareto stationary front learning. Empirically, MOSEL uncovers diverse and optimal Pareto frontiers in strongly conflicting settings (e.g., fairness-accuracy). Remarkably, even in weakly conflicting regimes such as multi-task learning, it consistently converges to solutions closer to the utopia point, outperforming both standard single-objective training and specialized multi-task learning methods. These results highlight the broader potential of a posteriori MOO learning as a pathway to efficiently learn more diverse and robust representations, ultimately improving generalization.

Using 2.5D particle-in-cell simulations, we investigate the parameters of a plasma wakefield accelerator (PWFA) for positrons and electrons in a layered plasma waveguide (LPW) in the presence of a transition layer (TL) between two layers of uniform plasma with different densities. The layered plasma is modeled as a combination of a near-wall tubular plasma of higher density and an on-axis plasma column of substantially lower density, separated by a TL in which the plasma density varies smoothly from the lower to the higher value according to a specified law. It is shown that using the PWFA parameters determined for a layered plasma waveguide without the transition layer changes the acceleration conditions when the TL is present, and that this change depends on the TL thickness. It is demonstrated that when the TL thickness exceeds the skin-layer depth in the tubular plasma, the positron acceleration rate decreases substantially. The methods are proposed for correcting the parameters of the plasma waveguide under consideration that compensate for the decrease in the energy gain of the accelerated bunches and ensure their stable transport.

Magnetism is a fundamental property of matter, typically arising from collective ordering of electronic spin and orbital angular momentum. Other quasiparticles interact with magnetism. However, they have not been considered its primary source. Here we show that phonons can generate a distinct form of magnetism through spontaneous ordering of their angular momentum. This phonomagnetism emerges from spin-phonon angular-momentum coupling, producing an effective exchange interaction between local ionic angular momenta. We show that the interaction strength depends on spin susceptibility, with a phonomagnetic phase predicted near a ferromagnetic critical point. Our theory also reveals phononic analogues of established spin models, including phonon Heisenberg, phonon $\Gamma$-model, and phonon Dzyaloshinskii-Moriya interactions. Combining first-principles phonon calculations with recent data from Raman spectroscopy, we identify VI$_3$ as a candidate phonomagnet, providing a microscopic framework for interpreting the weak magnetic anomaly reported above its bulk ferromagnetic transition. These findings open avenues for new platforms for information storage and manipulation on picosecond timescales.

Human-object interaction (HOI) detection aims to localize human-object pairs and recognize their interactions. Traditional supervised methods perform strongly but rely on task-specific training. Recent multimodal large language models (MLLMs) offer a promising route to training-free HOI detection through their broad visual-semantic knowledge and versatile perceptual and reasoning capabilities. However, existing approaches largely invoke these capabilities through loosely coordinated inference stages. This fragmented execution restricts the role of interaction hypotheses in guiding visual exploration, leaving key participants overlooked and local ambiguities unresolved. Furthermore, propagating early semantic assumptions through subsequent visual grounding and relation prediction induces self-reinforcing semantic circularity. To resolve these challenges, we propose HarnessHOI, a training-free framework that transforms passive MLLM inference into an active interaction-centric harness. Specifically, we introduce an interaction-guided perception mechanism that projects emerging interaction hypotheses back into the visual space to discover missing participants and refine ambiguous evidence through targeted observation. Furthermore, a relation-agnostic geometric adjudication module reconciles multi-source evidence to establish a unified spatial basis for grounded interaction reasoning across multiple actions and semantic roles. Extensive experiments on HICO-DET and V-COCO demonstrate that HarnessHOI achieves state-of-the-art performance among training-free methods, confirming the effectiveness of the proposed harness for complex interaction understanding. Code will be released upon publication.

To study the onset of scalarization, we investigate linear scalar perturbations around Kerr–Newman black holes in the Einstein-Chern-Simons-Maxwell-scalar theory with quadratic Chern–Simons (CS) coupling $\gamma$ ($\hat\gamma=\gamma/M^2$) and Maxwell coupling $\eta$. For this purpose, we employ hyperboloidal Rácz–Tóth evolution and a stationary axisymmetric zero-mode computation. The resulting thresholds are compared with the time-domain response at representative points. For $\hat\gamma=0$, we find two asymmetric threshold branches in $\eta$, while for $\hat\gamma=\eta$, the two branches have opposite signs and unequal magnitudes. For unequal CS and Maxwell couplings, the two real threshold branches merge as the spin or CS coupling increases.

We present an exact self-consistent two-component reformulation of the Dirac equation in the Weyl representation. By exploiting the diagonal structure of the kinetic operator and the purely scalar mass coupling in the Weyl basis, the four-component Dirac problem is reduced to an exact two-component form without introducing inverse-potential operators that arise in conventional energy-decoupling approaches. The resulting second-order differential equations are cast into an integral formulation and solved iteratively via convolution with the Helmholtz Green's function. We implement this formalism within an adaptive multiwavelet basis framework, providing rigorous, user-defined error control. Proof-of-concept calculations for simple atomic systems demonstrate the numerical stability, accuracy, and efficiency of the proposed algorithm.

Optical frequency combs with fast tunability are essential for integrated spectroscopy and sensing. However, achieving such functionality in the terahertz regime presents a significant challenge. This work introduces a novel approach that exploits supermode symmetry control in terahertz lasers. By electrically manipulating the gain ratio between coupled concentric ring lasers, we selectively excite symmetric and antisymmetric supermodes, allowing frequency tuning and spectral modulation across different frequency regimes and operation modes. Moreover, simultaneous lasing of both supermodes results in a reduced total integrated dispersion, widening the comb locking range. Supported by matrix formalism and numerical simulations, this work demonstrates the potential of symmetry control for enhanced electro-optic modulation, opening new avenues for on-chip spectroscopy and sensing applications.

We formulate a nonlocal spin-$1/2$ field theory in curved spacetime in which the form factor is a nonpolynomial entire function of the covariant Dirac operator. We derive the corresponding nonlocal Dirac equation and analyze its coupling to gravity and to a $\rm U(1)$ gauge field. As an application, we study the theory in the open chart of de Sitter spacetime and construct the associated nonlocal Bunch–Davies modes. The nonlocal dynamics can be mapped, branch by branch, onto the local Dirac equation with an effective mass $M_n$. For the exponential form factors considered here, the effective masses are expressed in terms of branches of the Lambert-$W$ function. We show that the Euclidean analyticity structure of the Bunch–Davies construction is preserved. On the physical positive-residue branch, the Bogoliubov coefficients, squeezing parameter, occupation number, and entanglement entropy acquire explicit nonlocal corrections, while the absence of physical fermionic supercurvature modes is preserved. The local theory is smoothly recovered in the limit $\Lambda\to\infty$.

Super-resolution and semantic segmentation are known to benefit one another, especially in the Earth observation context. However, learning both tasks in a joint model often requires both task annotations, which is impractical and expensive. In this paper, we study the multi-task partially supervised learning paradigm for both tasks, where each example is assumed to have only a single-task annotation. To that end, we examine two multi-task architectural variations, the sequential and shared variants, and then propose a hybrid variant and a re-projection loss to benefit from the shared representation and enforce image quality of super-resolution when training with semantic segmentation. Experiments show favorable results compared to the SOTA sequential variant. Source code will be published at https://github.com/lhoangan/munera.

We develop a combinatorial and Bayesian framework to quantify the information gained by increasing the local photon-number resolution of multiplexed detection systems based on superconducting nanowire single-photon detectors (SNSPDs). Motivated by recent studies that attribute photon-number resolving capabilities to a single SNSPD overcoming the typical on/off detection mechanism, we generalize the existing formulation for the binary devices by deriving the probability response tensor of multiplexed systems based on SNSPDs with an arbitrary number of resolvable states. We include finite efficiency through binomial loss convolution and, separately, assess the impact of dark counts. From this response tensor we derive a generalized N-photon detector fidelity, a Bayesian photon-number reconstruction, and the corresponding reconstruction fidelity. We particularize this formulation to study 2-, 3- and 4-state multiplexed systems with different numbers of detectors and show that increasing local photon-number resolution systematically improves both detector fidelity and Bayesian photon-number reconstruction. This substantially relaxes the quadratic scaling between detector number and resolvable photon-number required by purely binary multiplexed arrays. These results provide a quantitative value-of-information assessment for multi-state photon-number-resolving detector arrays.

Large language models (LLMs) have achieved substantial performance gains through increases in model size, training data, and computational resources. However, traditional scaling approaches produce diminishing returns, rising financial and environmental costs, and barriers to participation for researchers operating outside large industrial laboratories. This review examines the evolution of LLM scaling theory from empirical scaling laws to compute-optimal training, with particular emphasis on parameter efficiency, token utilization, data efficiency, and resource-constrained environments. Foundational work on scaling laws is synthesized alongside later research on compute-optimal training, data pruning, efficient architectures, quantization, low-rank adaptation, and edge-oriented optimization. The literature indicates a shift from scale maximization toward more deliberate allocation of parameters, tokens, compute, and hardware resources. At the same time, important empirical, theoretical, and methodological gaps remain regarding whether scaling principles established on enterprise-grade infrastructure generalize to smaller models and constrained computing environments. This review organizes these developments into a unified framework for resource-efficient LLM training and argues that future progress should evaluate efficiency not solely through model performance, but through the relationship among performance, parameter count, computational cost, token allocation, and hardware constraints.

It was recently proved that an embedded minimal annulus with free boundary in the unit ball $\Bbb B^3$ of the Euclidean space is, up to a rotation, the critical catenoid. We prove that an immersed, possibly branched, free boundary minimal annulus in $\Bbb B^3$ with embedded boundary components is necessarily free of branch points and is embedded. This shows the uniqueness of the critical catenoid holds under these weaker hypotheses. We then apply these results to prove that if $\Omega$ is a sufficiently regular annulus in $\Bbb S^2$ for which there exists a smooth function satisfying the overdetermined eigenvalue problem \begin{equation*} \begin{cases} \Delta u+ 2u = 0 \quad \text{on} \quad \Omega\\ \qquad \quad u=0 \quad \text{in}\quad\partial\Omega\\ \quad\,\,\,\, |\nabla u|=1 \quad \text{in}\quad \partial\Omega. \end{cases} \end{equation*} then $\Omega$ is, up to rotation, a rotational annulus with equatorial symmetry.

We establish global energy compactness for rough three-dimensional axisymmetric Euler flow without swirl and use it to construct hydrodynamic limits of collisional kinetic equations beyond finite-time Euler regularity breakdown. For relative vorticity in $L^1\cap L^p$, regular approximate Euler flows are globally compact in $C_tL_x^2$ for $p\ge4/3$, without a sign condition or any spatial-moment assumption. Below $4/3$, the same conclusion holds under uniform control of the absolute impulse. We identify the sharp radial dynamics governing this confinement: the absolute impulse is propagated without a sign condition for $p\ge5/3$, while for $1<p<5/3$ it is controlled under one-sided $L^{p^\sharp}$ integrability, where $p^\sharp= {4p} / (3p-1)$, and this exponent is sharp. In particular, for one-sign finite-impulse vorticity the resulting global energy compactness holds throughout the full range $p>1$. For the Coulomb Landau equation and a class of non-cutoff soft-potential Boltzmann equations, we construct strong solutions on lifespans $T^\varepsilon\to\infty$ realizing prescribed rough finite-energy Euler initial data and obtain energy-conserving axisymmetric Euler limits on every fixed time interval. If the corresponding regular Euler evolution loses regularity in finite time, the hydrodynamic fields converge to the regular Euler solution up to the singular time and have a common strong trace there, while subsequences yield global energy-conserving Euler continuations past that time.

We establish that the skew-symmetric bicharacter on the group of $\ell$-weights of a quantum affine algebra is invariant under Chari's braid group action. Furthermore, we define an extension of this bicharacter to the group generated by prefundamental $\ell$-weights and show that this invariance continues to hold under the extended braid group action of Frenkel-Hernandez. As immediate applications, we obtain braid group actions by automorphisms on the quantum tori associated respectively with the category of finite-dimensional representations of the quantum affine algebra and with the category $\mathcal{O}^{\mathfrak{b}}$ of its Borel subalgebra. In addition, we derive explicit formulas for quantum $QQ$-systems in all simply-laced types. Finally, we show that our extended bicharacter yields a quantization matrix compatible with the Hernandez-Leclerc cluster algebra categorified by $\mathcal{O}^{\mathfrak{b}}$ for all finite Dynkin types, extending Bittmann's construction in the simply-laced setting.

Inference-time steering offers cheap, fine-grained control over a language model's outputs by estimating a concept's representation in activation space and shifting activations towards it. Existing methods build these representations from activation averages over contrastive datasets. These averages incorporate unrelated concepts and noise, and are dominated by a few tokens, meaning the activation transport encodes token-level rather than thematic concepts. In this work, we steer towards examples that most express a concept thematically, rather than towards an expectation over all. We identify these examples using influence functions, which estimate how much each data point contributes to a model's representation of a concept. Unlike simple model activation similarity, they incorporate the curvature of the model's loss landscape, allowing them to capture concept-relevant relationships beyond superficial token-level similarity. We then propose influence-weighted activation transport, which uses optimal transport to steer activations of non-concept text towards those of concept text, weighting concept examples by their influence scores. We evaluate on toxicity suppression (Jigsaw), object-based concept induction (OneSec) and truthfulness induction (TruthfulQA), outperforming existing activation-transport baselines. We track capability after steering using perplexity and MMLU accuracy, finding that our method improves steering while largely preserving model quality. We further show that influence functions capture concept-relevant information that activation-based methods miss with the two approaches ranking data points significantly differently. Together, these results demonstrate the value of curvature-aware influence information for activation steering.

Let $d\geq2$ be an integer, and let $\operatorname{MPoly}^d$ be the moduli space of degree-$d$ polynomials. For every number field $K$, we prove that there exists $\epsilon_{K,d}>0$ such that \[ \left\lbrace\alpha\in\operatorname{MPoly}^d(K^{\mathrm{ab}})\colon h_{\mathrm{crit}}(\alpha)<\epsilon_{K,d}\right\rbrace=\left\lbrace\alpha\in\operatorname{MPoly}^d(K^{\mathrm{ab}})\colon h_{\mathrm{crit}}(\alpha)=0\right\rbrace \] is finite, where $h_{\mathrm{crit}}$ is the critical height. In particular, only finitely many $K^{\mathrm{ab}}$-rational points of $\operatorname{MPoly}^d$ are postcritically finite. The proof combines the correspondence method of Ji–Song–Xie with local ramification estimates and the uniform finiteness of intertwined classes in $\operatorname{MPoly}^d$. We also prove an analogous finiteness theorem for non-preperiodic marked points in non-isotrivial one-parameter polynomial families.

We study the position and shape of invasion fronts for spatially periodic Fisher–KPP equations on higher-dimensional lattices, starting from localized initial data. Periodic spectral geometry determines a directional logarithmic upper bound. When the linear growth rate is independent of the propagation coordinate, we prove the matching axial Bramson correction and identify the critical pulsating front approached by the solution. The selected profile includes the local periodic cell and a bounded phase whose derivative tends to zero. This description also determines the times at which prescribed levels reach distant lattice sites and the asymptotic time gaps between levels or neighboring cells. The proof combines a periodic transverse heat kernel, a scalar lower comparison, and critical-tail analysis. For growth rates varying in every coordinate, a positive Dirichlet quotient comparison reduces the matching lower bound to two explicit estimates on the linear seed and its accumulated nonlinear loss.

Aerosol particles play a critical role in air quality, human health, and climate processes, making their accurate characterization essential. One common way to represent aerosol populations is through particle number size distributions (PNSDs), which describe the concentration of particles across different mobility diameters. However, these distributions are often highly skewed, heavy-tailed, and multi-modal, posing significant challenges for conventional fitting approaches. In this study, we propose a robust parametric framework (AeroMiG) for modeling PNSDs using Inverse-Gamma mixture distributions. The method represents observed size distributions as a superposition of multiple Inverse-Gamma components, enabling flexible modeling of asymmetric and heavy-tailed structures that are not well captured by widely-used Log-Gaussian based approaches. The performance of the proposed method is evaluated using real-world data from three measurement stations. Results are assessed using multiple criteria, including predictive accuracies, computational efficiency, and a composite score summarizing overall model quality. Comparative analysis against a widely used Aerosol Multi-mode Log-Gaussian (AeroMG) demonstrates that the AeroMiG framework consistently achieves improved fitting accuracy, lower error, and faster computation. The proposed approach provides a reliable and computationally efficient solution for modeling complex aerosol size distributions, with potential applications in real-time air quality monitoring and data-driven environmental analysis.

This work presents two hardware-accelerated strategies for the upgrade of the CMS Level-1 Trigger (L1T) to target efficiency losses due to highly energetic radiating muons and unconventional physics signatures, such as Long-Lived Particles (LLPs). Operating directly on low-level detector hits, the first strategy implements a dedicated muon shower identification algorithm in the CMS barrel region that monitors hit multiplicities in the Drift Tube chambers to generate Shower Primitives. Executing with sub-bunch-crossing latency and ultra-low resource consumption, this algorithm successfully tags showered muons and provides essential context for downstream track finders. The second approach targets the barrel-endcap transition (overlap) region by evaluating Graph Neural Networks to reconstruct displaced muon tracks. Following a divided methodology that first isolates hardware feasibility from evolving physics refinements, an initial proxy model based on a GraphSAGE architecture is evaluated to validate its viability on FPGAs within strict latency limits. This implementation utilizes an INT8-PO2 quantization technique coupled with a data-driven bit-width optimization. By replacing DSP-heavy fixed-point multiplications with fast compile-time arithmetic bit-shifts and allocating the minimum required bits per signal, this hardware-software co-design reduces DSP utilization to $20\%$ and achieves a deterministic inference latency of just $19$ clock cycles. Together, these developments pave the way for triggering on non-standard muon signatures, such as those from LLP decays, at L1T.

The rapid evolution of video generation has narrowed the perceptual gap between authentic and synthetic videos, making generalizable AI-generated video detection increasingly challenging. Existing detectors predominantly rely on visual representations, leaving caption-derived textual semantics underexplored. Meanwhile, temporal regularity in fine-grained visual representations has received limited attention. We find that caption-derived textual representations provide complementary discriminative cues to global visual representations. Our analysis further reveals that AI-generated videos exhibit stronger temporal persistence and lower temporal variability, a pattern we term temporal over-regularity (TOR). Based on these findings, we propose MTOR with a multimodal branch and a TOR component. The multimodal branch integrates global visual and caption-derived textual representations, while the TOR component models temporal over-regularity at three levels: coarse inter-frame continuity, fine-grained token correspondence, and frame-to-video stability. Extensive evaluations on five benchmarks covering 46 generator variants demonstrate state-of-the-art overall performance against 16 representative baselines, while robustness experiments confirm strong resilience to twelve real-world video perturbations. Code and models will be released at https://github.com/hwang-cs-ime/MTOR.

Background: The Expected Value of Sample Information (EVSI) quantifies the value of collecting additional evidence to inform a health economic model. Existing EVSI methods typically assume idealized data collection mechanisms, most commonly randomized controlled trials (RCTs). However, in many realistic contexts, additional evidence comes from observational studies, which are subject to confounding or other bias. In this work, we define a methodology to calculate EVSI when additional data are observational and confounded. Methods: First, we define a simulation-based framework in which confounded observational data are generated through Inverse Target Trial Emulation (ITTE), a methodology that generates observational data with controlled levels of confounding starting from initial level data or prior information on population structure. Then, we apply inverse probability weighting (IPW) to obtain an adjusted summary statistic of the data targeting the corresponding randomized estimand. EVSI is finally computed using a regression based approach. Moreover, we propose a computationally efficient method to determine the sample size required to recover the EVSI achievable under an idealized randomized design. Results: We apply the methodology to two health economic models: a Normal Normal conjugate model and a chemotherapy treatment model combining a decision tree and Markov structure. We show that EVSI computed from observational data, even after adjustment, is lower than EVSI based on randomized data. The loss in EVSI increases with the level of induced confounding. Conclusions: This methodology extends EVSI when future evidence is expected to be observational and affected by confounding, enabling value-of-information analysis in more realistic and feasible data collection scenarios. It also provides a principled approach to sample size planning when observational evidence is anticipated.

We propose to use Quality-Diversity (QD) algorithms to solve robotic prehensile manipulation tasks in open-ended environments. Our approach enables the efficient discovery of a wide range robust grasp configurations, which serve as reliable starting points for generating diverse prehensile manipulation trajectories on articulated objects. The resulting diversity in manipulation behaviors enhances generalization and adaptability, enabling effective deployment continuously evolving open-world settings.

A pseudo-Riemannian manifold is called geodesic orbit (GO) if every geodesic is an orbit of a one-parameter group of isometries. For Riemannian $H$-type groups, the GO property was completely characterized by A. Kaplan and C. Riehm in terms of the dimension of the centre and the structure of the underlying Clifford module. We solve the corresponding problem for pseudo $H$-type groups $N_{r,s}$, that is, 2-step nilpotent Lie groups whose Lie algebra $\mathfrak n_{r,s}=\mathfrak z\oplus\mathfrak v$ is built from an admissible module $\mathfrak v$ of the Clifford algebra $\mathrm{Cl}(\mathbb R^{r,s})$, endowed with the left-invariant metric whose restriction to the centre is indefinite ($s\ge 1$). We prove that $N_{r,s}$ is naturally reductive if and only if $(r,s)\in\{(0,1),(1,2)\}$, and then it is GO for every admissible module. We also prove that $N_{3,4}$ is GO, but not naturally reductive, exactly when $\mathfrak v$ is a minimal admissible module, and that in all remaining cases $N_{r,s}$ is not GO. Thus, in contrast with the Riemannian situation, where a seven-dimensional centre admits GO metrics on isotypic modules of dimensions 8, 16 and 24, the exceptional pseudo-Riemannian case is GO only for the smallest possible module. For $N_{3,4}(\mathfrak v_{\min})$ we explicitly construct the one-parameter isometry groups generating all geodesics, including the null ones, and describe the affine family of their generators.

Administrative boundary reforms in China alter the assignment of local authority and growth incentives without changing physical geography. We study two reforms with opposite institutional content—county-to-district (centralization: a county absorbed into a municipality as an urban district) and county-to-city (empowerment: a county upgraded to a county-level city)—using a firm-year pollution panel (1998–2013) and a border-pair research design that compares firms on opposite sides of the same county border. Centralizing reform lowers local NOx intensity by 14.8% through cleaner production, without shrinking output or employment; empowering reform raises COD, dust, and wastewater levels by 87–95% through dirty growth. The local abatement is not net abatement: NOx rises in same-prefecture neighbor counties (+0.77, p<0.01), reallocation follows administrative boundaries rather than geographic distance, is confirmed in satellite ambient data, and is not driven by wind or water transport. The reallocation operates through incumbent neighbor firms becoming dirtier, not through firm relocation or new entry. A back-of-the-envelope welfare calculation yields a statistically insignificant welfare gain from the only significant abatement (NOx, \textyen 47 million/yr, CI crossing zero) and a damage of -\textyen 2.3 billion/yr from county-to-city reform. These findings show that boundary reforms reshape the spatial configuration of pollution across administrative boundaries rather than reducing it.

Braess paradox originates when latency at Wardrop equilibrium in traffic networks decreases because of removing edges. The graph-theoretic property of networks suffering from the Braess paradox was called vulnerability by Roughgarden in 2006; it was then characterized and algorithmically checked both for undirected and for directed nets. In this paper, we provide a decremental algorithm of linear amortized complexity to check vulnerability for dynamically evolving networks. The basic idea of our dynamic algorithm is to use a static algorithm that marks some edges as irrelevant for the vulnerability of the graph and ignore those edges for all subsequent runs of the decremental procedure. To get a linear amortized cost for every edge remotion, we also provide a new version of such a static algorithm that improves its complexity from O(n m^2) to O(m^2), that is in turn aligned with the cost of the best state-of-the-art static algorithm for vulnerability.

We study empirical processes of observable extreme generators in stationary Poisson-Laguerre tessellations, with each generator weighted either by one or by the volume of its cell. Under moment assumptions that allow heavy-tailed marks, we prove functional central limit theorems as the observation window grows, on compact intervals where the mark law has a bounded density. We also obtain covariance asymptotics with a surface-order error and quantitative univariate and multivariate Gaussian approximation.

We investigate the effects of quenched disorder on two-dimensional electron solids using Hartree-Fock theory and classical molecular dynamics. Increasing disorder weakens crystalline correlations, broadens the structure factor, and reduces the measured coherence length toward the pristine lattice spacing. This evolution is consistent with the Imry-Ma and Larkin picture: random pinning disrupts translational order on sufficiently long scales, and stronger disorder reduces the distance over which crystalline organization survives. Classical simulations further show that stronger pinning raises the temperature for particle rearrangement even after crystalline order has disappeared. Disorder thus suppresses structural coherence while stabilizing an amorphous solid against thermal rearrangement.

We study scalaron quasinormal modes and quasi-bound states of Schwarzschild black holes directly within the Bondi-Sachs formulation of metric $f(R)$ gravity, combining frequency-domain analyses with characteristic time evolution. Starting from the linearised characteristic equations, we obtain the massive scalar master equation and show that the tensor master equation takes the same form as in general relativity. For the analytic models considered here, the linear scalar spectra depend on the $f(R)$ theory only through the dimensionless combination $\mu M$. We compute both quasinormal and quasi-bound spectral branches using Leaver's continued-fraction method and compare them with independent literature benchmarks. To complement the frequency domain analysis, we evolve the scalaron perturbation equations using a fourth order convergent characteristic scheme. Applying matrix-pencil stability scans to the resulting waveforms, we independently recover fundamental quasinormal frequencies consistent with the continued-fraction results within the measured extraction-stability scales. Finally, for a selected finite-mass case, we demonstrate the extraction of both oscillation frequencies and damping rates of the fundamental quasinormal mode and quasi-bound state from separate time windows of the same waveform at one extraction radius, without imposing either spectral branch at the finite outer boundary.