Papers

Automated service systems increasingly generate algorithmic operational decisions that shape how services are delivered. However, these decisions reach end-users only through frontline workers who carry them out in real-world settings. During this process, automated decisions can diverge from user expectations, surfacing as friction at the service encounter. We propose Frontstage Mediation Work as a preliminary analytic lens for examining the often invisible labor through which frontline workers anticipate and manage such misalignments between algorithmic decisions and user expectations. Drawing on a qualitative case study of an On-Demand Ride-Pooling service, we identify four recurring practices through which drivers sustain the service encounter when frictions arise. Such labor remains absorbed into routine operations, leaving no trace in performance metrics, system logs, or formal job descriptions. This paper contributes to worker-centered HCI scholarship by illustrating how automated services shift onto frontline workers the responsibility of managing the interactional consequences of system-level decisions.

Intrusion Detection Systems (IDS) struggle with imbalanced datasets like NSL-KDD, especially in detecting rare R2L and U2R attacks. This work describes a systematically optimized and explainable framework using a CNN-Transformer architecture to improve performance on highly imbalanced data. We decided to use XGBoost for feature selection and Focal Loss as the main mechanism for minority class learning. We used Optuna for end-to-end hyrate=0.00042, batch size=256, and Focal Loss {\gamma}), using a thoughtful data splitting process to prevent data leakage. Critically, our ablation studies pointed out that while Focal Loss (optimally at {\gamma} = 1.5) substantially enhanced minority recall, adding oversampling methods like SMOTE caused precision degradation via over-correction. Our final model, using only tuned Focal Loss, achieved 98.73% overall accuracy on the NSL-KDD test data, with significantly balanced F1-scores of 84.63% (R2L) and 69.66% (U2R) over baseline approaches. Additionally, we performed SHAP analysis to understand model predictions, understand salient features (e.g., service http, logged in), and explain the persistent U2R precision challenge due to feature overlap. This study presents a comprehensive pipeline for imbalanced IDS, providing methodological insights, and demonstrates that tuned Focal Loss can be a sufficient and effective strategy for class balancing.

Transiting Warm Jupiters orbiting bright stars are scarcely known, but a substantial number of these planets have been uncovered by the TESS mission. In this work we report the discovery and confirmation of a transiting Warm Jupiter orbiting the bright ($V$ = 8.70) subgiant star HD 715 (TOI-6893). HD 715 b was originally identified from a single transit event in TESS Sector 70. We acquired two seasons of radial velocity follow-up using the PFS, TRES, and MINERVA-Australis spectrographs, which clearly detect the planetary orbit. At $M_*=1.60\pm0.04 M_\odot$, HD 715 thus becomes one of the most massive stars known to host a transiting Warm Jupiter. From a joint fit to the TESS transit and our RVs, we find that HD 715 b has a $56.40^{+0.19}_{-0.23}$ day orbital period and measure a precise mass and radius of $0.635\pm0.056 M_J$ and $1.07\pm0.04 R_J$, rendering it a relatively typical member of the Warm Jupiter population. The moderate orbital eccentricity of $0.21^{+0.07}_{-0.08}$ contrasts with the higher eccentricities of other similar Warm Jupiters orbiting evolved stars, indicating a greater diversity in orbital eccentricities than previously known. The absence of additional transits in preceding TESS observations in Sectors 42 and 43 allows us to further refine the range of possible orbital periods to $[56.01, 56.35]$ days. Though the ephemeris uncertainty remains substantial, future photometric observations may nonetheless allow for recovery of the transit of HD 715 b.

We study fair division of indivisible items under eligibility constraints: each item may be assigned only to a subset of eligible agents. This model coincides with restricted assignment in scheduling, with agents as machines and items as jobs, and captures settings where work or resources must be split among parties with different capabilities. Eligibility makes the usual fairness benchmarks inadequate, as agents should not be measured against items they could never receive. We therefore use eligibility-aware versions of proportionality and envy-freeness, where an agent's proportional share splits each item equally among its eligible agents, and envy comparisons discount items the agents are not eligible for. Our main finding is a sharp chore–goods separation for proportionality. We focus on identical valuations: under arbitrary valuations, no online algorithm can guarantee even an $o(m)$-approximation for chores. For chores, Prop1 is achievable offline, whereas exact proportionality and proportionality up to any item are not. For online chores, the optimal approximation factor for Prop1 for deterministic algorithms is $\Theta(\sqrt m)$, with the lower bound holding even for unit-size jobs. For goods, Prop1 is again achievable offline, but the tight online bound drops to $\Theta(\log m)$. For envy-based fairness, the picture is different. Goods and chores are equivalent under eligibility-aware envy notions, while the choice of notion is decisive. Strong EF1 may fail to exist even offline and cannot be approximated online. In contrast, weak EF1 can be maintained online, while weak EFX separates offline from online. Together, these results give a systematic study of online proportional and envy-based fairness in restricted assignment and show which fairness guarantees remain achievable under eligibility constraints.

This paper presents the HINTT system submitted to the 2nd Challenge and Workshop on Multilingual Conversational Speech Language Model (MLC-SLM). We address multilingual speaker-attributed ASR, where systems must determine who spoke when and what was spoken. We investigate two modeling strategies for this problem: a cascaded pipeline that combines speaker diarization with speech-LLM-based ASR, and a unified speech LLM that directly generates speaker labels, timestamps, and transcriptions. Our final submission is based on the cascaded pipeline, consisting of a fine-tuned DiariZen diarization model, a fine-tuned Qwen3-ASR model, and LLM-based generative error correction. For comparison, we also fine-tune VibeVoice-ASR as a unified model using the same official training data. All task-specific fine-tuning and model selection are performed using only the official MLC-SLM data, without external data or pseudo-labels. Experimental results demonstrate that the cascaded system remains more reliable under the MLC-SLM Task 1 conditions, while unified speech LLMs offer a promising direction for future speaker-attributed ASR.

A central goal in quantum information science is understanding which physical states can be converted into others. The ideal scenario would be a totally ordered system, where every state can be ranked by a single value, and a higher-ranked state can always be converted into a lower-ranked one. Such an ordering would provide a complete description of state convertibility, making the resource structure maximally simple to characterize and compare. We prove that such a simple ordering is fundamentally impossible for any quantum system with dimension larger than 2. The underlying properties of quantum mechanics guarantee that some states will always be incomparable. Furthermore, we show this is a uniquely quantum phenomenon by constructing classical systems of any size that can be perfectly ranked. This establishes a rigorous and striking divide between quantum and classical state transformations.

Supervised fine-tuning can substantially improve the downstream utility of large language models (LLMs) but may compromise their safety. Existing safety-preserving methods constrain downstream updates using safety-related parameters or subspaces, but mainly focus on safety preservation rather than joint safety and utility enhancement, lack a theoretical characterization of the optimal safety-related subspace and safety-preserving task update, and typically rely on a static safety subspace that may become outdated during fine-tuning. To address these limitations, we propose ASCENT, a downstream fine-tuning framework for safety–utility co-enhancement through first-order optimal safety-aware periodic calibration and task optimization. We model safety as a function of LLM parameters $S(\theta)$ and use its first-order approximation to characterize safety changes under parameter updates. Under a fixed rank and Frobenius-norm budget, we prove that the update constructed from the top-$r$ singular components of the safety-function gradient maximizes the estimated safety change, and use it for periodic calibration to preserve and improve safety. We further derive a unique safety-preserving task update that stays close to the original task update while penalizing negative effects on the estimated safety change. ASCENT alternates these optimal task and calibration updates to jointly enhance safety and utility. Experiments across multiple LLM families and downstream tasks show that ASCENT improves downstream utility by up to 20.3% and reduces attack success rate by up to 35.5%, achieving state-of-the-art safety and utility across all evaluated settings. Our code is available at https://github.com/ZJU-LLM-Safety/ASCENT.

Nanoparticle radiosensitization depends on interconnected NP properties, yet geometry, elemental composition, surface coating, and localization are typically studied separately, with their combined biological impact rarely assessed in one framework. Drawing on TOPAS Monte Carlo simulations from PhD research, we consolidated four investigations of metallic nanorods, mainly gold nanorods (AuNRs), examining these parameters across closely linked studies. The findings reveal that each parameter can influence over a distinct spatial range. Nanorod geometry primarily alters the spectrum and distribution of low-energy Auger-Meitner electrons emitted within tens of nanometers. Elemental composition induces significant near-field dose enhancement that rapidly attenuates and becomes negligible beyond roughly 1 um. Surface coatings reduce the fluence of secondary electrons below 3.5 keV, absorbed within about 150 nm of the surface, lowering physical dose contribution by 1-7%. Since coatings predominantly remove low-energy electrons, radiolytic yield changes through a shift in the electron energy spectrum, not directly following dose response. At the vascular scale, the enhanced dose region extends to approximately 10 um from the vessel wall, meaning NP localization determines where the biological target sits relative to radiation sources. Collectively, no single design parameter serves as a universal predictor of radiosensitization. Translating physical radioenhancement into biological effect depends on whether the target lies within each parameter's effective range. Furthermore, charged-particle equilibrium (CPE) correction, applied where beam confinement necessitated it, substantially alters calculated dose enhancement ratios. Consequently, scoring distance and CPE-corrected values should be explicitly reported to ensure comparability across studies.

In this article, we use a family of distributions developed by Topp-Leone to create a novel lifetime two-parameter distribution. The Topp-Leone XLindley distribution is the name used for it. We study several types of statistical and mathematical characteristics of this distribution, such as reliability functions, moments, moment-generating function, quantile function, and the Renyi entropy. For the purpose of estimating parameters under the proposed distribution model, this simulation study is conducted to evaluate the performance of the maximum likelihood estimates for the parameters of the Topp-Leone XLindley distribution. A comprehensive simulation investigation is carried out to evaluate the performance of the proposed methods. Furthermore, a practical data set of bank customer waiting times has been investigated for the purpose of illustration.

We prove a local Hartman-Grobman theorem at zero for the unforced Navier-Stokes equation on the three-dimensional torus in the strong topology of $H^s$ for every $s>7/2$. More precisely, we construct a homeomorphism between open positively invariant neighborhoods of zero that conjugates the Navier-Stokes semiflow to the heat semigroup. We develop an infinite-dimensional boundary-time scheme built on a nested hierarchy of exact asymptotic dynamical factors. Their fiber geometry yields Sobolev block coordinates with autonomous finite-dimensional factor systems, on which boundary-time topological conjugacies can be constructed at every finite level and then assembled on a common neighborhood into the desired $H^s$ conjugacy. The infinite-dimensional assembly relies on new spectral estimates for the divergence-free transport structure, which control derivative loss and the blockwise tails of the finite-level conjugacies and their inverses without an additional spectral-gap assumption.

We develop a highest weight framework and construct explicit finite dimensional irreducible modules for the type AIII (quasi split) quantum symmetric pair subalgebras $U_q'(\mathfrak{gl}_n^\theta)$. In Part 1, we provide a uniform construction of Cartan subalgebras and Verma modules for both subfamilies. Our construction recovers Letzter's Cartan subalgebra in one subfamily and gives a new Cartan subalgebra in the other. We construct a weak triangular decomposition of $U_q'(\mathfrak{gl}_n^\theta)$ via the Letzter map. Then we define Verma modules as induced modules and prove that finite dimensional simple $U_q'(\mathfrak{gl}_n^\theta)$-modules are quotients of our Verma modules. In Part 2, we define a diagrammatic category $Web^B$ that controls polynomial representations of $U_q'(\mathfrak{gl}_n^\theta)$. We prove a multiplicity-free decomposition of the quantum exterior powers $\bigwedge^k\mathbb{V}$ as a $U_q'(\mathfrak{gl}_n^\theta)$-module, using a explicit eigenspace decomposition of certain dot morphisms in $Web^B$. As applications, we deduce the multiplicity-freeness of the involved anti-spherical Hecke module. For $\bigwedge^k\mathbb{V}$, we explicitly determine the highest weight vectors and their weights, thereby determining the isomorphism classes of their irreducible summands. Finally, we determine the kernel of the $Web^B$-action on certain finite dimensional representation category of $U_q'(\mathfrak{gl}_n^\theta)$.

This work presents a prototype implementation of a noncoherent multiple-input multiple-output (MIMO) system based on Grassmannian signalling, where neither the transmitter nor receiver requires channel state information (CSI), operating at a carrier frequency of 15 GHz in the frequency range 3 (FR3) upper mid-band envisioned for 6G. To the best of our knowledge, this is the first prototype implementation of Grassmannian signaling reported in the literature beyond sub-6 GHz bands. We identify the motivation of applying Grassmannian signaling, a form of unitary space-time modulation (USTM), in this band for wireless and mobile applications with low spectral-efficiency requirements but strict reliability and power-efficiency demands. With this idea in mind, we present a system design integrated with orthogonal frequency-division multiplexing (OFDM) that addresses the challenges associated with the higher carrier frequencies of FR3 while providing reliable operation across a broad range of scenarios, including urban and rural environments as well as vehicular communications. We then present the prototype implementation of the proposed design, based on software-defined radio (SDR) devices integrated with radio-frequency (RF) mixers, filters, and power amplifiers to ensure adequate operation in the targeted band. Over-the-air (OTA) experiments based on symbol error rate (SER) measurements under non-line-of-sight (NLoS) and line-of-sight (LoS) scenarios, and across different data rates, show good agreement with software simulations, thereby validating the prototype implementation.

Automatic assessment of communication quality in counseling conversations is bottlenecked by data: expert-rated corpora are small and expensive to grow. Therefore we study cross-domain prediction of the expert's overall impression across three German simulated-counseling corpora from one university (two general-practice medical, one parent-teacher; $n=195$ sessions). Features are construct scores that four small LLM judges assign from the transcript, drawn from the experts' own rating instruments, plus a compact nonverbal block; all models are open-weight and run locally, preserving data protection. Trained on the other two corpora, this stack predicts the held-out corpus at Spearman $\rho = .533$; in-domain training on the same sessions falls about a fifth short ($.420$; paired gap $+.113$, 95% CI $[+.067, +.157]$), and the gap holds at matched training size and on identical test sessions. The signal lies in what is said: the four judges alone transfer at $.523$, and the nonverbal block adds $+.01$ (CI includes zero). Asking a judge for instrument constructs instead of generic dialogue qualities raises its transfer by about 30% ($.32$ to $.42$); restoring speaker attribution where per-speaker audio was lost is worth $+.08$ to $+.10$ on that fold. Where a program's own ratings are scarce, other programs' data is a viable training set and language carries it.

The detailed Monte Carlo simulation of optical calorimeters is dominated by the generation and tracking of scintillation and Cherenkov photons, which in standard Geant4 ray-tracing increases the CPU time per electromagnetic shower by up to three orders of magnitude compared to simulating only the energy deposition. At the luminosity and granularity foreseen for the LHCb ECAL Upgrade II, this cost becomes prohibitive for both detector R&D and large-scale sample production. The optical photons cannot simply be discarded, because the time resolution of a few tens of picoseconds targeted by the upgrade to mitigate pile-up is driven by their arrival-time distribution at the photo-detectors. This paper presents HybridMC, a fast and detailed simulation framework for optical calorimeters that decouples shower development from optical transport. Energy deposition is handled by a standard Geant4 simulation without optical photon propagation, while light transport is reproduced by a parametrised summary Monte Carlo step based on probability density functions extracted from a one-time full ray-tracing optical calibration. A Monte Carlo-to-Monte Carlo validation shows good statistical agreement with full ray-tracing for all module types of the upgraded LHCb ECAL, from individual photon arrival times up to the reconstructed energy and time resolutions. The CPU time is reduced by more than two orders of magnitude, with module-dependent acceleration factors from 150 to 480. A comparison to test-beam data from DESY II and the CERN SPS shows that the framework reproduces the momentum dependence of the measured energy and time resolutions of a real prototype, up to a small residual constant term. HybridMC retains the microscopic optical detail not accessible to flash simulators based on generative models, at a CPU cost compatible with large-scale production on standard computing infrastructure.

This paper addresses an extension of the pickup and delivery problem with time windows and multiple stacks (PDPTWMS). We call the new extension the pickup and delivery problem with time windows and requests split across multiple stacks (PDPTWRSMS). Applications of the PDPTWRSMS arise in less-than-truckload transportation systems where loads are packed as pallets or other standardised containers. In the PDPTWRSMS, the vehicle's loading space is divided into compartments of finite capacity. Loading and unloading in each compartment must obey the last-in-first-out (LIFO) policy. Moreover, a request can be split among all compartments, and a single request can also exceed a single compartment's capacity. Thus, each request is only limited to the vehicle's capacity. Computational experiments show that applying this extension can lead to substantial routing cost reductions (the total distance travelled and/or the number of vehicles used), and, compared with versions of the PDPTWMS where compartments are not restricted to a LIFO policy, splitting is worth more once the vehicle has three or more compartments, though less when it has two deep ones. We apply a modified version of the recently proposed method of fragments. Results confirm that this approach can solve many PDPTWRSMS instances and significantly outperforms the current state-of-the-art method for the PDPTWMS.

We construct smooth Einstein metrics with positive scalar curvature on $\#_k({\mathbb S^2}\times {\mathbb S^2})$ and on $k\mathbb{CP}^2\#(2k+1)\overline{\mathbb{CP}}{}^{2}$, for $k\geqslant2$. In particular, we show that positive Einstein metrics exist on infinitely many homeomorphism types of closed $4$-manifolds. The first family shows that any simply connected spin $4$-manifold carrying metrics with positive scalar curvature is homeomorphic to an Einstein $4$-manifold. Wick rotation gives associated Lorentzian static axisymmetric Einstein metrics with positive cosmological constant and arbitrarily many horizons with the same surface gravity. The metrics are obtained by varying the cone angle $2\pi\beta$ of a prescribed singularity along a torus, starting from the product metric on $\mathbb S^2\times\mathbb S^2$ and from the symmetric Kähler-Einstein metric on $\mathbb{CP}^2\#3\overline{\mathbb{CP}}{}^2$. At $\beta=\frac{2}{k+1}$, changing the period of the angle around the torus removes the cone singularity and changes the topology. For both families, as $k\to+\infty$, the metrics collapse along the torus orbits to an explicit $2$-dimensional disk, whose boundary has Hausdorff dimension $\frac{4}{3}$. At the scale of curvature, they converge to complete periodic Ricci-flat metrics with Kasner asymptotics: the Euclidean Myers-Korotkin-Nicolai metric for the first family, and a new one for the second.

The origin of supermassive black holes, especially those already present at high redshift, remains an open problem. We propose that their seeds are primordial black holes formed from rare overdense regions of conventional thermal DM. A spectator scalar with two degenerate vacua gives the DM different masses, and hence different freeze-out abundances, in the two vacua. Inflationary fluctuations populate the vacuum containing the heavier relic only in exponentially rare patches. As the Universe cools, the density contrast of these patches grows sufficiently large for the patches to collapse into PBHs after horizon re-entry. We derive the resulting PBH mass function and show that it is sharply concentrated near a lower cutoff while reproducing the mass scale and abundance required for SMBH seeds. Since the overdensities originate from a rare, non-Gaussian population rather than from an enhancement of nearly Gaussian primordial curvature perturbations, the usual CMB spectral-distortion constraint does not directly apply. The construction also correlates the seed population with the dark-sector spectrum and the inflation scale, providing complementary avenues for testing the scenario.

We construct a smooth embedded open annulus in Euclidean three-space with strictly negative Gaussian curvature and injective Gauss map that contains a simple closed asymptotic curve of zero normal self-linking. This gives an affirmative answer to Problem 1.1 of Ghomi and Raffaelli. The construction starts with two simple planar curves whose tangent vectors are positive multiples of one another. Their mixed integral is made zero by rounding a pair of polygons with one free length parameter.

Low-rank adaptation (LoRA) is widely used as a parameter-efficient fine-tuning technique for pre-trained deep neural networks, which approximates the weight update via full fine-tuning by a low-rank matrix $BA^\top$. This parameterization leads to the equivalence relation $(B, A) \sim (BG^{-1}, AG^\top)$ for any invertible matrix $G$ because $BA^\top = BG^{-1}(AG^\top)^\top$ and thus both pairs yield the same loss value. This relation induces a quotient manifold where matrices $(BG^{-1}, AG^\top)$ for all $G$ are identified, eliminating redundant directions along which the loss value remains unchanged. To respect the geometry of this manifold, the original search space is endowed with a Riemannian metric that is invariant under the equivalence relation. Such a metric induces preconditioning at each gradient step and ensures that each weight update via LoRA changes the loss value, leading to efficient optimization. In this paper, we propose a new Riemannian metric that is specifically tailored to LoRA to close the gap to full fine-tuning at the weight level. We theoretically show that LoRA with our preconditioning induced by this metric satisfies the following two properties at each iteration: (i) The weight update follows the direction closest to the gradient of full fine-tuning within the subspace of first-order weight changes allowed by the LoRA parameterization. (ii) The updated weight matrix is closer in Frobenius norm to that of full fine-tuning than the updated weight matrices of LoRA with conventional preconditioning and without preconditioning. These theoretical insights suggest that our preconditioning makes LoRA better approximate full fine-tuning, thereby leading to more efficient optimization. Experiments show the effectiveness and efficiency of our preconditioning for LoRA on fine-tuning tasks with language and vision domains.

LLM agents often lack the operational knowledge to act reliably in new environments, as they must discover specific tool behaviors or environment conventions on their own. Without memory of past attempts, they repeat the same mistakes across tasks, leading to more task failures and longer trajectories. To address this, agentic systems typically rely on human-written guidelines or on procedural memory built from training tasks and an oracle verifier, both of which require prior knowledge of the environment. We present DAEDALUS, a method for bootstrapping reusable agent memory from self-generated practice without existing tasks or oracle verifiers. DAEDALUS pairs two agents: an explorer that interacts with the environment to generate challenging yet solvable tasks, and a solver that attempts them. A heuristic is derived from each solver failure and accepted only after the solver repeatedly succeeds with that heuristic in context. These outcomes also provide feedback for the explorer to refine the difficulty of future tasks. Accepted heuristics are then consolidated into a memory bank for test-time use. Across AppWorld, $\tau^2$-bench, and AutomationBench, DAEDALUS improves mean success rates by up to 15.9 points and pass^5 by up to 2.2x over a no-memory baseline, and is competitive with methods using training tasks, at a lower inference cost than most. We show that performance gains already emerge with a small exploration budget, and that its heuristics also benefit agents from other model families. Our ablations further reveal that solver traces provide the key information needed to derive effective heuristics, while factorizing early discoveries makes exploration more cost-efficient. Beyond memory construction, we find that the tasks generated by DAEDALUS can serve as a proxy for benchmark tasks when ranking models by performance. Code and artifacts: www.github.com/illuin-tech/daedalus.

Achromaticity is one of the most demanding functionalities achievable with a metasurface because it requires reproducing the same optical response over a finite spectral bandwidth rather than at a single wavelength. Although previous studies established aperture-bandwidth limits based on temporal-delay considerations, the role of optical performance on these limits has remained unresolved. Here, we develop a rigorous theoretical framework that explicitly highlights how the optical performance sets the true aperture-bandwidth limit of single interface achromatic metasurfaces, quantifying the fundamental limits through the Strehl ratio. We identified two distinct operating regimes. In the first, the achromaticity conditions are exactly satisfied, guaranteeing diffraction-limited performance across the entire bandwidth. Surprisingly, the maximum achievable aperture in this regime is essentially independent of bandwidth and is instead governed by the ability of practical meta-atoms to simultaneously realize the required phase and dispersion. In the second regime, relaxing the required optical performance allows larger apertures at the cost of a fundamental trade-off between aperture, bandwidth, and optical performance. Numerical validation using a large library of high-transmission GaN meta-atoms confirms the theory and reveals the transition between these regimes. Our framework establishes the fundamental limits of achromatic metasurfaces, providing practical design rules for balancing aperture, bandwidth, and optical performance of achromatic metasurfaces.

Early sepsis warning from ICU records can be cast as a structure-preserving prediction problem. A model needs to detect deterioration from irregular measurements while keeping each alert connected to the physiological signals that support it. Many temporal models fuse clinical variables into a patient-level representation, supporting scalar risk prediction but weakening the structure needed for clinical decomposition. We present SepsisLens, which preserves variable-indexed temporal states until risk composition. Observation-aware representations encode each variable's dynamics and measurement history, while a shared temporal encoder models each trajectory without collapsing the variable axis. The StructuredRiskHead composes multi-horizon risk from explicit variable-level and organ-level components. We evaluate SepsisLens on three public ICU cohorts and one private-hospital cohort under a common pre-onset protocol. SepsisLens achieves strong discrimination on all four cohorts and lower alert burden at matched event recall on MIMIC-IV. Structural ablations support the design, while input-side masking shows that the ranked components reflect variables with greater influence on prediction.

We investigate scalar induced gravitational waves (SIGWs) at third-order during a matter-dominated (MD) era. We classify the sources into four parts—built from three first-order scalars, and from a first-order scalar times the second-order tensor, vector, and scalar perturbations—and obtain all four kernel functions analytically in closed form. Unlike their radiation-dominated counterparts, the MD kernel functions grow secularly, translating into a rapid growth of the spectrum, $\Omega^{(3)}_{\mathrm{GW}}\propto x^{6}$ versus $\Omega^{(2)}_{\mathrm{GW}}\propto x^{2}$ at $x\equiv k\eta\gg1$. We derive the third-order power spectrum as a five-fold integral for a general primordial spectrum, and evaluate it numerically for a monochromatic one, where all contributions from the second-order vector source vanish identically. As an application, we consider the early MD era driven by primordial black holes (PBHs) before their evaporation. Requiring that the induced gravitational waves do not backreact on the cosmological perturbations, $\int\mathrm{d}\ln k\, \Omega^{\mathrm{tot}}_{\mathrm{GW}}<1$ then bounds the initial PBH abundance as $\Omega_{\mathrm{PBH},f}\lesssim4.6\times10^{-7}\, (m_{\mathrm{PBH}}/10^{7}\,\mathrm{g})^{-0.48}$, about three orders of magnitude stronger than the second-order-only bound. Our results indicate that higher-order SIGWs inevitably become nonperturbative during a sufficiently long MD era, with important implications for scenarios involving an early matter domination.

We propose a new formula relating the rationalized Drinfeld associators obtained from the Knizhnik–Zamolodchikov equation to the quantum $6j$-symbols of the algebra $\mathcal{U}_q(\mathfrak{g})$: $ U_{q} = U \cdot \Phi^{KZ}_{\mathbb{Q}} $. We prove this relation in the case of the $\mathfrak{sl}_N$ Lie algebra for the tensor cube of the fundamental and the first symmetric representation. We also check it perturbatively for the $\mathfrak{sl}_N$ algebra for representations $[2]\otimes [1] \otimes [1]$, $[2]\otimes [2] \otimes [1]$, the non-trivial cases with multiplicities — $[2,1]^{\otimes 3}$, $[2,2]^{\otimes 3}$, and in the case of the $\mathfrak{so}_5$ Lie algebra for the tensor cube of the fundamental representation.

Ensuring safety in autonomous driving requires continuous map maintenance supported by reliable quality indicators. In this context, it is crucial to identify when and where map updates should be triggered, for instance through crowdsourced data, and under which conditions a new map compilation should be deployed. This paper focuses on effective metrics for assessing the quality of vector maps and guiding such decisions. We present a new metric called GOSPAM designed to measure map discrepancies in terms of location errors, existence, and completeness. Through detailed simulations on both point and polyline feature maps, we analyze its sensitivity to common map degradation such as bias, false positives, false negatives, and coordinate errors. The results demonstrate that GOSPAM offers a unified and interpretable measure that effectively captures various forms of map deviation, making it a strong candidate for map quality assessment in automotive applications.

In this paper, we study the existence problem of completely independent spanning trees (CIST) in chordal graphs through appropriate hypergraph representations and their panchromatic and bipanchromatic colorings. First, we disprove a conjecture stating an exact relationship between the panchromatic number, the bipanchromatic number, and the minimum number of unique colors in an optimal panchromatic coloring of a hypergraph. Then, by relating CIST to panchromatic and bipanchromatic colorings of the associated hypergraphs, we derive structural conditions for their existence in chordal graphs and specifically strictly chordal graphs.

High precision continuous localization is a fundamental challenge for automated guided vehicles (AGVs) operating in electromagnetically sensitive industrial environments. To mitigate the continuity limitations of conventional discrete localization methods and the accuracy limitations of existing continuous localization methods, this paper proposes a continuous AGV localization method that integrates interferometric fiber Bragg grating (FBG) array sensing with a hybrid deep learning framework. We first design a shallowly buried interferometric FBG array sensing track to enable the stable acquisition of high-quality signals. The received sensing signals are subsequently parameterized and preprocessed in a task-oriented manner, in which a multichannel joint representation is constructed using a sliding time window. A hybrid deep learning framework is developed to estimate AGV position from multichannel FBG response signals by jointly exploiting inter-channel spatial correlation and temporal continuity in the measurement sequence. Experimental results show that the proposed method achieves a mean absolute error (MAE) of less than 2 cm under complex industrial conditions, proving its effectiveness and robustness in practical AGV localization scenarios.

High-entropy alloy (HEA) nanoparticles offer chemically diverse catalytic sites, but identifying optimal compositions across their large design space remains challenging. We developed a framework combining machine-learning energy prediction, thermodynamic Monte Carlo sampling, Bayesian optimization, and kinetic Monte Carlo (kMC) simulations to screen 85-atom nanoparticle models across a nine-element composition space for the hydrogen evolution reaction. Static optimization based on adsorption energies selected Pt-rich, low-component compositions, whereas kMC-based optimization ranked the multielement composition Au$_{33}$Co$_{20}$Cu$_{6}$Pd$_{26}$ highest among the evaluated candidates. Cycle-resolved analysis further showed hydrogen adsorption at Co/Pd-rich environments, migration through intermediate-binding sites, and frequent Heyrovsky desorption at weaker-binding, Au-containing environments. This pathway-level evidence supports a multisite reaction picture and illustrates how incorporating kinetics can complement static adsorption-energy screening.

Let $\Omega=G/K$ be an irreducible bounded symmetric domain other than the unit disk, and let $S$ be its Shilov boundary. We study the two-parameter family of Shilov–Poisson transforms between homogeneous line bundles over $S$ and $\Omega$. For a transform $F$ of $L^2$ boundary data, we determine what finite Euclidean regularity up to the full boundary $\partial\Omega$ forces in the interior. Outside five explicit exceptional parameter families, sufficiently high finite regularity forces $F=0$. In each exceptional family, the conclusion is an interior covariant differential equation. In the scalar Poisson–Szegő case, it yields a finite-order Graham type theorem: the stated boundary regularity forces a Bergman-harmonic function to be pluriharmonic.

A spin one-half driven only by total-spin operators generates a renewal process in which one interevent time $n^\ast$ is forbidden while all others occur. No classical hidden Markov model reproduces it with fewer than $1+\sqrt{n^\ast}$ states, an unbounded separation whose only resource is one relative phase of $\pi$. With $K$ amplitudes the same interference yields $K(K{+}1)/2$ decay channels, and for positive weights and generic rates the classical cost is exactly $K(K{+}1)/2$ states. Both are realized by a depth-three circuit whose parameter count does not grow with the register.