Context: Along with developing Deep learning (DL) models, larger datasets and more complex model structures are applied, leading to rising computing resources and energy consumption, which is an alert that green DL models should receive more attention. Objective: This paper focuses on a novel view to analyze DL energy consumption: the effect of hyperparameters on the energy cost of DL models. Method: Our approach involves using mutation operators to simulate how practitioners adjust hyperparameters, such as epochs and learning rates. We train the original and mutated models separately and gather energy information and run-time performance metrics. Moreover, we focus on the parallel scenario where multiple DL models are trained in parallel. Results: To examine the effect of hyperparameters on energy consumption, we conducted extensive experiments on five real-world DL models. The results show that (1) many hyperparameters studied have a (positive or negative) correlation with energy consumption, (2) adjusting hyperparameters can make DL models greener, i.e., lead to less energy consumption without performance damage, and (3) in a parallel environment, energy consumption becomes more susceptible to change. Conclusions: We suggest that hyperparameters need more attention in developing DL models, as appropriately adjusting hyperparameters would cause green DL models.

Pulse-shape discrimination (PSD) in high-purity germanium (HPGe) detectors is central to rare-event searches such as neutrinoless double-beta decay (0vBB), yet conventional approaches compress each waveform into a small set of summary parameters, potentially discarding information in the full time series that is relevant for classification. We benchmark transformer-based models that operate directly on digitised waveforms using the Majorana Demonstrator AI/ML data release. Models are trained to reproduce the collaboration-provided accept/reject labels for four standard PSD cuts and to regress calibrated energy. We compare supervised training from scratch, masked autoencoder (MAE) self-supervised pre-training followed by fine-tuning, and a feature-based gradient-boosted decision tree (GBDT) baseline. Transformers outperform GBDT across all PSD targets, with the largest gains on the most challenging labels and on the combined PSD-pass definition. MAE pre-training improves sample efficiency, reducing labelled-data requirements by factors of 2-4 in low-label regimes. For energy regression, both transformer variants show a small common underestimation on the test split, while fine-tuning modestly narrows the residual distribution. These results motivate follow-up studies of robustness across detectors and operating conditions and of performance near QBB.

We consider two-dimensional Coulomb systems for which the coincidence set contains an outpost in the form of a suitable Jordan curve. We study asymptotics for correlations along the union of the outpost and the outer boundary of the droplet. These correlations turn out to have a universal character and are given in terms of the reproducing kernel for a certain Hilbert space of analytic functions, generalizing the Szegő type edge correlations obtained recently by Ameur and Cronvall. There are several additional results, for example on the effect of insertion of an exterior point charge in the presence of an outpost.

Providing quantitative interpretation of coherent nonlinear microscopy images, such as third-harmonic generation (THG), is generally hampered by the complex phase-matching conditions, especially in the presence of sample linear heterogeneity. We recently presented a numerical pipeline using the finite-difference time-domain (FDTD) method to take this heterogeneity into account. However, due to software restrictions, we only considered nonlinear materials with diagonal nonlinear susceptibilities. We now expand the recently developed FDTD approach to model nonlinear microscopy for anisotropic materials that obey Kleinman Symmetry, organized in layers along the optical axis, and validate our simulations on well-described geometries.

We study thin films that host coexisting collinear $d$-wave altermagnetic and superconducting orders in the presence of an external magnetic field that fully penetrates the films. We use the Ginzburg-Landau functional to analyze the response of the films to magnetic fields and an in-plane supercurrent. We demonstrate that the interplay between superconductivity and altermagnetism induces characteristic fourfold anisotropies in the critical temperature, parallel critical field, and critical current density of the films. These results establish experimentally accessible signatures of altermagnetism in superconducting films and in superconductor/altermagnetic insulator heterostructures.

The composition $\mathcal{F} \circ \mathcal{G}$ of two combinatorial classes $\mathcal{F}$ and $\mathcal{G}$ is a standard combinatorial construction and translates into the composition $F(G(z))$ of their corresponding counting generating functions. Such a composition is called critical if $G(ρ_G) = ρ_F$, where $ρ_F$ and $ρ_G$ denote the corresponding radii of convergences of $F$ and $G$, respectively. In this case, both the singular behaviours of $F$ and $G$ influence that of $F\circ G$. Such critical decomposition schemes appear quite frequently in the context of map enumeration. For example by using the block-decomposition one has $M(z) = B(z(1+M(z))^2)$ and $ρ_B = ρ_M (1+M(ρ_M))^2$, where $M(z)$ denotes the generating series of all rooted planar maps and $B(y)$ the generating series of $2$-connected rooted planar maps. This can be extended to multivariate generating functions by taking several statistics into account, for example face counts. Since critical composition schemes show (usually) a condensation phenomenon -- in the above situation this means that there is giant $2$-connected block of linear size and linearly many small blocks -- it is very plausible that statistical properties on $2$-connected maps transfer to corresponding properties of all maps and back. The purpose of the present paper is to make this precise on the level of the singular structure of the corresponding multivariate generating functions. In particular we show that moving $3/2$-singularities transfer. Since such kind of singularities are closely related to central limit theorems of the corresponding statistics this methods provides also a kind of transfer of central limit theorems. Actually this method is quite flexible and is applied to a variety of face and pattern counting statistics in map enumeration.

We provide a framework to theoretically describe long-range energy transfer in single and twisted two-dimensional hyperbolic slabs. We demonstrate that phonon polaritons (PhPs, quantum superpositions of photons and lattice vibrations in polar dielectrics) can mediate and enhance room-temperature energy transfer at ranges far exceeding those of conventional mid-infrared (MIR) platforms, and with extreme directionality. This is because the dipole-dipole interaction potential energy diverges along the asymptotes of the real-space hyperbolic opening angle. Our findings allow us to extend classical and quantum interactions between dipoles, typically strictly confined to the near-field, beyond several free-space MIR wavelengths. We use $α$-MoO$_3$ as a representative material, but this mechanism is not limited to the MIR: it is general to anisotropic media across the whole electromagnetic spectrum.

This paper presents a mathematical analysis of a one-dimensional model of turbulence based on a stochastic generalized Constantin-Lax-Majda-DeGregorio (gCLMG) equation. We focus on the specific case where the nonlinearity in the equation allows the existence of the anomalous enstrophy cascade, which is an inviscid conserved quantity, and some effective energy estimates for mathematical analysis. The existence of an invariant measure in the attractor is proved via the classical Krylov-Bogoliubov argument. The uniqueness of the measure and exponential mixing are proved under a sufficiently large viscosity condition, in which the nonlocal structure of the nonlinear term plays a prominent role. The construction of this invariant measure is the first step towards a theoretical understanding of turbulent phenomena that cause anomalous cascades in the zero viscous limit, viewed from the dynamical systems theory.

In this paper, we study the posets of classes of subgroups of finite group having same set of orders of elements. We show that this poset is a chain only in the case of p-groups and moreover, we characterize all finite groups for which this poset is C2, the chain with two elements. We also show that this poset forms a lattice in the case of finite cyclic and dihedral groups and give a characterization when this lattice is distributive and modular.

A finite non-degenerate set-theoretic solution $(X,r)$ of the Yang-Baxter equation gives rise to a structure skew brace $B(X,r)$ that is a $λ_f$-skew brace, i.e. every element has finitely many $λ$-images, and whose additive group is $FC$. This motivates the study of finiteness conditions on skew braces. We first study the general class of $λ_f$ skew braces and the subclass where the additive group is $FC$, showing that these properties share a resemblance to finite conjugacy, having an analog of the $FC$-center and several analogous structural results. Furthermore, by passing through the structure skew brace of a solution, this property measures whether elements are contained in a finite decomposition factor, identifying a class of infinite solutions that may exhibit similar properties to finite ones. Finally, we show that for a sub skew brace where both groups have finite index, both indices need to coincide and that such a sub skew brace contains a strong left ideal of finite index.

We study high-dimensional drift estimation for Lévy-driven Ornstein--Uhlenbeck processes based on discrete observations. Assuming sparsity of the drift matrix, we analyze Lasso and Slope estimators constructed from approximate likelihoods and derive sharp nonasymptotic oracle inequalities. Our bounds disentangle the contributions of discretization error and stochastic fluctuations, and establish minimax optimal convergence rates under suitable choices of tuning parameters in a high-frequency regime. We further quantify the sample complexity required to attain these rates depending on the Lévy noise. The results extend the theory of high-dimensional statistics for stochastic processes to a substantially broader class of noise mechanisms, in particular pure jump processes. They also demonstrate that Lasso and Slope remain competitive for jump-driven systems, providing practical guidance for inference in applications where Lévy processes are a natural modeling choice.

Let $(X, \mfA,P)$, $(Y, \mfB,Q)$ be two arbitrary probability spaces and $¶:=\{(\mfA,P_y):y\in{Y}\}$ be a regular conditional probability on $\mfA$ with respect to $Q$. Denote by $R$ the skew product of $P$ and $Q$ determined by $\{P_y:y\in{Y}\}$ on the product $σ$-algebra $\mfA\otimes\mfB$ and by $\wh{R}$ its completion. I prove that a process $\{ξ_y:y\in{Y}\}$ possesses an equivalent $\wh{R}$-measurable version if and only if it is measurable with respect to a certain particular $σ$-algebra, larger than $\mfA\otimes\mfB$ and uniquely determined by $¶$. It is known that not every process possesses an equivalent measurable version (cf. \cite[§19.5]{St}). My approach is essentially different from earlier trials. It reverts to \cite[Theorem 3]{ta1}, where Talagrand proved existence of an equivalent separable version of a measurable process (in case of $R=P\times{Q}$), provided $Y$ is endowed with a separable pseudometric. The theorem is a strong generalization of \cite[Theorem 6.1]{smm} and \cite[Theorem 5.1]{mms1} where it was proved only that a suitable class of liftings transfer a measurable process into a measurable process.

We look at the threshold effects in neutral and charged Drell-Yan production, Higgs boson production with a massive vector boson, and Higgs production in bottom quark annihilation at the Large Hadron Collider (LHC), up to the third order in QCD. Using third-order soft-virtual (SV) results and the universal properties of threshold logarithms, we find the process-dependent coefficients and improve the accuracy by including large threshold logarithms up to next-to-next-to-next-to-leading logarithmic (N$^3$LL) order and matched with the latest N$^3$LO results. We also show numerical results for the invariant mass distributions and total production cross sections for these processes. Our findings show that the theoretical scale uncertainties, which are about $0.4\%$ at N$^3$LO in fixed-order calculations, decrease to less than $0.1\%$ at N$^3$LO+N$^3$LL after SV threshold resummation in the high invariant mass region.

The secrecy performance of continuous-aperture array (CAPA)-based wiretap channels in terms of secrecy rate and secrecy outage probability (SOP) is analyzed. First, the system models of CAPA systems with maximum-ratio transmission under a Rayleigh fading channel are established, and approximate probability density functions for the legitimate user Bob's signal-to-noise ratio (SNR) and the eavesdropper Eve's SNR are derived using Mercer's theorem and Landau's eigenvalue theorem. Three scenarios are considered, including a single Eve, multiple independent Eves, and multiple collaborative Eves. Next, the expressions of the secrecy rate and SOP under these three scenarios are derived, and the high-SNR slope, high-SNR power offset, diversity order, and array gain in Bob's high-SNR region are obtained. It is then theoretically proven that, in all three scenarios, the CAPA system achieves the same high-SNR slope and the same diversity order, with the latter being equal to the spatial degrees of freedom. Moreover, the CAPA system with a single Eve has the smallest high-SNR offset and the highest array gain, whereas the CAPA system with multiple collaborative Eves exhibits the largest high-SNR offset and the lowest array gain. Finally, the theoretical analyses of secrecy rate, SOP, high-SNR performance are validated by the simulation results, and a higher secrecy rate and a lower SOP are achieved by the CAPA systems compared to the spatially-discrete array systems with half-wavelength antenna spacing.

While provably secure steganography provides strong concealment by ensuring stego carriers are indistinguishable from natural samples, such systems remain vulnerable to real-world edit errors (e.g., insertions, deletions, substitutions) because their decoding depends on perfect synchronization and lacks error-correcting capability. To bridge this gap, we propose Alkaid, a provably secure steganographic scheme resilient to edit errors via distance-constrained encoding. The key innovation integrates the minimum distance decoding principle directly into the encoding process by enforcing a strict lower bound on the edit distance between codewords of different messages. Specifically, if two candidate codewords violate this bound, they are merged to represent the same message, thereby guaranteeing reliable recovery. While maintaining provable security, we theoretically prove that Alkaid offers deterministic robustness against bounded errors. To implement this scheme efficiently, we adopt block-wise and batch processing. Extensive experiments demonstrate that Alkaid achieves decoding success rates of 99\% to 100\% across diverse error channels, delivers a payload of 0.2 bits per token for high embedding capacity, and maintains an encoding speed of 6.72 bits per second, significantly surpassing state-of-the-art (SOTA) methods in robustness, capacity, and efficiency.

This study investigates the lightest and second-lightest Higgs bosons at around 95 GeV (which is only a hypothetical scenario) and 125 GeV respectively within the CP-violating next to minimum B-L supersymmetric model(NB-LSSM). In the NB-LSSM, the CP violation in the Higgs potential leads to the mixing mass terms between the CP-even and CP-odd Higgs fields. Thus, one has to consider a $(10 \times 10)$-dimensional mass matrix for the neutral Higgs boson. These potential mixing effects may lead to drastic variations on the neutral Higgs boson masses. Besides, the neutral Higgs bosons predicted in the NB-LSSM may strongly mix with one another, thereby significantly modifying the couplings of the Higgs bosons to the fermions or gauge bosons. It is found that the specific parameters $g_{YB}$, $\tanβ$, $T_κ$, $T_λ, \cdot\cdot\cdot$ and CP-violating phases $θ_{1,2,3,4,6,7,8}$ in the NB-LSSM affect the theoretical predictions on the Higgs boson mass and corresponding signal strengthes significantly. And the theoretical predictions on the signal strengthes of SM-like Higgs decay channels and excess signals at around 95 GeV are fitted well to the observed experimental data.

Thermodynamics establishes that information acquired through measurement can be converted into work, as exemplified by Maxwell's demon and Szilard engines. Most experimental realizations of information engines, however, implicitly assume Markovian environments, in which information exchanged with the surroundings is irreversibly lost. Many physical systems instead exhibit environmental memory, with hidden degrees of freedom retaining correlations with the system's past and giving rise to non Markovian dynamics. Whether and how such concealed memory can be harnessed as a thermodynamic resource has remained an open question. Here we experimentally demonstrate work extraction from environmental memory. Using time resolved measurements on an optically trapped Brownian particle in equilibrium, we implement a time delayed double measurement protocol that retrieves information via backflow from hidden bath degrees of freedom. We show that this information backflow alters relaxation dynamics, can be quantified independently of initial state effects, and when appropriately exploited enhances work extraction. Notably, we identify regimes in which the extracted work exceeds the energy stored in the observable degree of freedom alone. Our results establish environmental memory as an experimentally accessible thermodynamic resource and reveal how non Markovian dynamics can be systematically explored to improve the performance of information engines operating in time-correlated environments.

Accurate and robust wireless localization is a key enabler for a wide range of mobile computing applications. Fingerprint-based localization using channel state information (CSI) has attracted significant attention due to its high accuracy and compatibility with existing communication infrastructures. However, traditional similarity-based fingerprinting methods suffer from high computational complexity and limited scalability in high-dimensional CSI spaces, while purely learning-based approaches fail to explicitly exploit correlations among reference fingerprints during inference. To address these challenges, this paper proposes a unified retrieval-assisted fingerprinting localization framework that tightly integrates similarity-based and learning-based paradigms. Specifically, channel charting is employed to project high-dimensional CSI into a low-dimensional latent space, enabling efficient and scalable retrieval of locally correlated reference points (RPs). Building upon the retrieved RPs, a graph attention network (GAT) is designed to explicitly model inter-sample correlations between the query CSI and its associated references, allowing adaptive and geometry-aware feature aggregation for accurate position estimation. Extensive experiments conducted on both real-world indoor and ray-tracing simulated outdoor scenarios demonstrate that the proposed method consistently outperforms state-of-the-art similarity-based and learning-based localization approaches.

We provide a method to systematically construct vector fields for which the dynamics display transitions corresponding to a desired hierarchical connection structure. This structure is given as a finite set of directed graphs $\mathbf{G}_1,\dotsc,\mathbf{G}_N$ (the lower level), together with another digraph $\mathbfΓ$ on $N$ vertices (the top level). The dynamic realizations of $\mathbf{G}_1,\dotsc,\mathbf{G}_N$ are heteroclinic networks and they can be thought of as individual connection patterns on a given set of states. Edges in $\mathbfΓ$ correspond to transitions between these different patterns. In our construction, the connections given through $\mathbfΓ$ are not heteroclinic, but excitable with zero threshold. This describes a dynamical transition between two invariant sets where every $δ$-neighborhood of the first set contains an initial condition with $ω$-limit in the second set. Thus, we prove a theorem that allows the systematic creation of hierarchical networks that are excitable on the top level, and heteroclinic on the lower level. Our results modify and extend the simplex realization method by Ashwin & Postlethwaite.

In this article, we construct new families of Ramanujan complexes with local structure distinct from all previously known examples. Our approach is based on unitary groups over number fields, more specifically on what we call super-definite unitary groups, that is definite unitary groups that are anisotropic modulo their center at a finite place. These arise naturally as groups of units in central division algebras with involution of the second kind. Our first main result gives a general construction of infinite families of Ramanujan complexes associated with a super-definite unitary group $G$ over a totally real number field and a finite place $v_0$. The structure of the resulting complex is governed by the type of the Bruhat-Tits building at $v_0$. It includes new examples of type $A_n$ when $v_0$ is split, and novel families of type ${}^2\!A'_n$, ${}^2 \! A''_n$ (with $n$ even), $B$-$C_n$, ${}^2 \! B$-$C_n$ and $C$-$BC_n$ in the non-split case. This construction works uniformly across all ranks. Since much of the motivation for constructing expander complexes comes from computer science, we investigate the algorithmic explicitness of our construction in the latter part of the paper, and provide an example in rank 5 where it becomes fully explicit. In particular, this example yields golden gates for the real Lie group $PU(5)$.

Border bases are traditionally restricted to 0-dimensional ideals due to the finiteness of the underlying order ideal. In this paper we extend the theory to homogeneous ideals of positive Krull dimension by introducing homogeneous border bases, defined relative to an infinite order ideal. Moreover, we provide two characterizations of these bases: one via border reductors and, most notably, one in terms of formal multiplication matrices. Although the latter condition a priori requires verification in infinitely many degrees, we prove that it is sufficient to check only finitely many of them, thereby obtaining an effective criterion.

Resonant chiral dielectric metasurfaces can support circular dichroism exceeding that of natural materials, but their small dissipative losses simultaneously limit the maximization of circular dichroism, which inherently relies on absorption. Importantly, while the condition for optimal circular dichroism in resonant structures can be rigorously formulated based on the concept of critical coupling, controlling the amount of absorption experimentally, and ideally tuning it to the optimal value post-fabrication, remains elusive. Here, we experimentally tailor the dissipative losses of chiral bilayer dielectric metasurfaces post-fabrication using energetic ion beam irradiation. Specifically, we study the transmission characteristics of C4-symmetric chiral metasurface consisting of silicon nanocuboid arrays embedded in silica glass using polarization-resolved spectroscopy. We enhance the circular dichroism from 0.70 in the pristine, unirradiated metasurface to 0.85 after irradiation. Our experimental results are complemented by numerical simulations allowing us to retrieve the refractive index changes induced by the ion beam irradiation in the constituent materials of the metasurface. Our work offers a new approach to globally maximize optical chirality in engineered nanostructures, paving the way towards chiral emission and advanced polarization control applications

This study investigates the use of machine learning based mesh adaptivity, specifically mesh movement methods (UM2N), with depth integrated non-hydrostatic shallow water models. Motivation for this comes from the need for models which balance efficiency and accuracy for use in probabilistic coastal hazard assessment. Implementations are built on the discontinuous Galerkin finite-element (DG-FE) based software, Thetis, which leverages the partial differential equation (PDE) framework Firedrake for automated code generation. Verification on benchmark test cases and validation against laboratory measurements of coastal hazards, focusing on tsunami propagation, run-up, and inundation is performed. In these tests, the UM2N-driven meshes help resolve key non-hydrostatic dynamics and yield numerical solutions in close agreement with reference computations and measured data. Numerical results indicate that the UM2N surrogate based approach significantly accelerates conventional mesh movement techniques and has high robustness over long integration periods and under strongly nonlinear wave conditions.

We present precise results for the $ZH$ production cross-section and invariant mass distribution at the LHC, taking into account the effects of the leading and sub-leading soft gluons. We improve both quark-initiated and gluon-initiated subprocesses through threshold resummation within the QCD framework.

As most of the companions in the double neutron star systems should be normal pulsars, the Fast Folding Algorithm (FFA), which is suitable for finding these long spin period pulsars, was used to search their possible radio signals. A time domain resampling code PYSOLATOR was used to maximize the available data length by removing the orbital modulation. We collected and processed 272.2 hours observational data taken by the Five-hundred-meter Aperture Spherical radio Telescope (FAST) for the 13 double neutron star systems in its sky. The signal-to-noise ratios of known pulsar signals are obviously improved by this search method, including the detection of a faint pulsar signal which only saw by folding the data. Unfortunately, no companion signals were found among all the 197962 candidates. Geodetic precession of the orbit could enhance detectability in future observations.

Advances in quantum computing threaten digital communication security by undermining the foundations of current public-key cryptography through Shor's quantum algorithm. This has driven the development of Post-Quantum Cryptography (PQC), a new set of algorithms resistant to quantum attacks. While NIST has standardised several PQC schemes, challenges remain in their adoption. This paper introduces the PQC-LEO framework, a benchmarking suite designed to automate the evaluation of PQC computational and networking performance across x86 and ARM architectures. A proof-of-concept evaluation was conducted to demonstrate the framework's capabilities and highlight its application in supporting ongoing research on the adoption of PQC algorithms. The results show that there is a greater performance reduction in implementing PQC methods with higher security on ARM architectures than on the x86 architecture.

The figure of merit $ρλ$ the product of resistivity and mean free path (MFP) evaluated from first-principles calculations, is widely adopted to screen promising interconnect metals with high electrical conductivity at ultranarrow dimensions. However, the $ρλ$ has been calculated without addressing the validity of the assumption that the MFP is independent of the wavevector $\mathbf{k}$. Here, we assess the validity of the constant MFP approximation, by estimating the $\mathbf{k}$-dependent MFPs for (an)isotropic elemental metals, with explicit treatment of electron-phonon interactions. Additionally, we verify the validity of the constant relaxation time approximation (CRTA) for resistivity calculations. We show that both the constant MFP approximation and CRTA are reasonable even for highly anisotropic Fermi surfaces. Our results support the practical use of those approximations in transport studies, where explicit electron-phonon calculations are not feasible.

Lattice vibrations carrying angular momentum, known as chiral phonons, have emerged as a promising route to control and understand complex material properties, yet their deterministic manipulation remains largely unexplored. Here we demonstrate electric-field switching of phonon angular momentum in the technologically relevant ferroelectric BaTiO3. Using circularly dichroic resonant inelastic X-ray scattering (CD-RIXS) at the oxygen K edge, we directly probe the phonon angular momentum and compare the measured dichroism with first-principles predictions of phonon-mode chirality. We find excellent agreement, revealing a momentum-dependent circular-dichroism contrast that exhibits a reversible gyroelectric effect, stable for at least 15 hours. Our results establish a robust mechanism for non-volatile control of chiral phonons and point towards new opportunities for phonon-based information and energy technologies.

We study Higgs-boson production in association with a top-quark pair ($t\bar{t}H$) at hadron colliders and present the first matching of next-to-next-to-leading order (NNLO) QCD corrections to parton showers using the MiNNLOPS method. For the two-loop amplitude, we employ two established approximations, based on the soft Higgs-boson and high-energy limits, respectively. For the first time, we also construct the latter in full colour and propose a pointwise combination of the two approximations across phase space. By assigning a conservative uncertainty estimate, which remains well below the perturbative uncertainties, we ensure robust and reliable differential predictions, explicitly validated at the one-loop level. Apart from the two-loop amplitude, all remaining ingredients of the MiNNLOPS calculation are included exactly. After thorough validation, we present a series of phenomenological results illustrating the impact of NNLO corrections and parton-shower effects. We consider fiducial predictions for the Higgs-boson decay into photons and include off-shell top-quark decays with tree-level spin correlations in both the dilepton and semileptonic channels. Our $t\bar{t}H$ MiNNLOPS generator is publicly available within the POWHEG framework.

In this note, we extend the definition of $p$-biharmonic and bi-$p$-harmonic maps between two Riemannian manifolds and explore some of their properties.