Papers

We study the no-flux problem for a taxis-type migration-consumption system with logistic damping in a bounded planar domain, where the motility function \(\phi\) vanishes at zero and degenerates algebraically of order \(\alpha\ge1\). We prove uniform boundedness in \(L^\infty(\Omega)\times W^{1,\infty}(\Omega)\) for every fixed damping coefficient \(b>0\), without any smallness assumption on the initial data, thus improving the previous \(L^2\)-boundedness results. For \(\alpha>1\), we further prove that the solution converges exponentially to the spatially homogeneous state \((a/b,0)\), with an explicit admissible rate. The proof combines a weighted Gagliardo–Nirenberg type inequality adapted to the degeneracy with the dissipative structure of the signal equation. Therefore logistic damping drives the system toward homogeneity, despite the loss of uniform elliptic regularization.

Robot learning has undergone significant advancements in the domain of compliant control and robotic assembly. The development of novel methodologies has transcended the capacity to address discrete tasks, such as grasping or inserting, and has evolved to encompass the comprehensive construction of entire assemblies. Conventional benchmarks, which are predominantly focused on single tasks, are no longer adequate to address the complexity and interconnected nature of modern manufacturing processes. Consequently, the necessity for benchmarks to encompass a broader spectrum of assembly and subassembly steps is becoming apparent. In light of these developments, we propose LABIT, an industrial robotic assembly benchmark for long-horizon manipulation. This benchmark encompasses DIN-norm assembly tasks, including screwing, placing, inserting, and snapping, under both high and low tolerance conditions. The successive nature of these assembly steps implies that prior errors can influence subsequent manipulation. Consequently, this benchmark is particularly relevant for assessing the robustness of intelligent robotic systems. The proposed benchmarking protocols and novel performance metric, based on a taxonomy of assembly complexity in conjunction with the comprehensive list of parts and their corresponding CAD models, serve as essential tools for research and development in the field. To provide a baseline, we evaluate the benchmark using traditional force-position control and report reliability and score metrics per primitive, subtask, assembly category, and subassembly. Accompanying data is available at https://irp-hka.github.io/robotic-assembly-benchmark/.

We study environment scaling as a way to increase agent-benchmark difficulty while holding core task objectives fixed. Our framework, Recursive Environment Scaling (RES), transforms existing benchmark environments while preserving their objectives and completion requirements. A proposer changes information access, resource relations, and execution dependencies; execution-grounded checks validate the resulting candidate, and accepted changes are inherited across rounds. This produces related environments for paired evaluation and training-trajectory collection. Experiments on 294 initial tasks from three benchmarks show lower task success for all six evaluation models, with decreases of 57.14–77.14 percentage points on Claw-Eval-Live. With equal trajectory counts and identical optimization settings, fine-tuning on RES trajectories outperforms original-trajectory SFT in all nine model–benchmark comparisons by 7.00–14.61 percentage points. Ablations identify recursive inheritance and dependency composition as important contributors to increased difficulty. Together, these results support environment scaling for both agent evaluation and training.

Guided sampling with sparse observations and PDE residuals has recently been combined with neural operator priors to solve partial differential equations in function space, their natural setting. We pair such a prior with pseudo-bootstrap sequential Monte Carlo (pBS SMC), a guided sampling framework that accommodates proposals without tractable densities, and introduce a new proposal by adapting the DPM++ algorithm to the guidance setting. Across five benchmark PDE systems this proposal attains the best error–time trade-off at every step count tested, and wrapping it in pBS gives the lowest error on most inverse-problem settings, with the largest gains when few sampling steps are used.

The fusion of artificial intelligence (AI) and space technologies is extending spaceborne intelligence beyond onboard payloads toward a global infrastructure, here termed Space AI infrastructure (SpaceAI-Infra). Its promise rests on two distinctive advantages of space systems: abundant solar energy in orbit and the global reach of satellite constellations. Realizing this promise hinges on a fundamental systems question: How should satellite constellations be designed and orchestrated to deliver data-center-scale AI services competitive with their terrestrial counterparts? This article develops a roadmap-oriented answer. We first chart the evolution of SpaceAI-Infra through three progressive stages: onboard AI, Space Edge AI, and Space Cloud AI. By revisiting the space-constrained fundamentals for end-to-end token services, we derive enabling design principles and translate them into an integrated space edge–cloud architecture. The communication-centric space edge comprises free-flying LEO constellations that provide user-proximate computing and form the access and backhaul fabric for cloud AI services. Complementing the edge, the computing-centric space cloud comprises formation-flying satellite constellations organized as space data centers in dawn–dusk sun-synchronous orbits. A case study demonstrates that the proposed architecture can support data-center-scale foundation-model inference for millions of concurrent users worldwide.

In 1965, Chowla conjectured that the values of the Liouville function become asymptotically uncorrelated at consecutive natural numbers. Recently, the two-point case is resolved by OpenAI. In 2026, Charamaras and Richter conjectured a generalization of the Chowla conjecture to the asymptotic independence of the consecutive prime Omega functions. In this article, we prove the two-point case of their conjecture. As a consequence, we obtain a dynamical generalization of the two-point Chowla conjecture.

We present a self-contained, algorithmic construction of almost invariant subspaces [Panati et al. (2003), De Nittis and Panati (2010), Freund and Teufel (2016)] for Hofstadter-like Hamiltonians $H_b $ acting on $\ell^2(\mathbb{Z}^2,\mathbb{C}^n)$. More precisely, for rational values of $b_0/(2\pi)$ we describe an iterative procedure that takes a “locally gapped" projection $P_{0}$ (possibly topologically non-trivial) which commutes with $H_{b_0}$, into a projection $P_{\varepsilon}$ that commutes with $H_{b_0+\varepsilon}$ up to an error of order $\varepsilon^\infty$. Our construction insures that the Chern number of $P_{\varepsilon}$ equals the Chern number of $P_0$ as long as $P_\varepsilon$ exists. Moreover, its integrated density of states varies linearly with $\varepsilon$ (the “Středa formula"). We also show that any spatially localized Parseval frame of $P_{0}$ can be mapped into to a localized Parseval frame of $P_{\varepsilon}$. We choose to work with a discrete model in order to keep the technicalities to a minimum level, with no need for advanced pseudo-differential calculus. Nevertheless, we are confident that most of our results can be extended to (unbounded) magnetic (pseudo)differential operators, but the technical burden would be much heavier. This is left for future work.

We study the following simultaneous embedding problem: Given k graphs on the same vertex set with the same pairwise intersection, do they admit planar drawings that coincide on their shared graph? We establish fixed-parameter tractability with respect to the vertex integrity of the union graph (that is, the union of the input graphs) plus k, resolving an open question and generalizing the known algorithm parameterized by vertex cover number plus k. Our main technical contribution on the way towards this result is the extension of the recently introduced Ramsey pruning technique beyond ordering problems. Finally, we complement our tractability result with a lower bound that establishes NP-hardness when both k and the pathwidth of the union graph are bounded by constants, ruling out a generalization of our main result to path- or treewidth.

Multi-step retrosynthesis planning is often considered solved once a route to purchasable starting materials is found, but such a route should also be practical in real-world chemistry. Recent LLM-based planners formulate this process as sequential decision-making over an AND-OR graph, yet typically expand a single molecule per turn and are optimized for solvability alone, leaving alternative routes uncompared across the multiple objectives that determine practicality before the search budget is committed. We introduce RetroPlanner, an LLM-based planner that jointly compares multiple routes to steer the search toward higher synthesizability, as measured by reaction plausibility, round-trip consistency, and starting-material price. RetroPlanner is trained to reason over chemical evidence: structural progress from the target molecule, bond changes and reaction classes, where candidate routes diverge from one another, and how they rank across multiple objectives. Experiments on multi-step retrosynthesis benchmarks demonstrate that RetroPlanner recovers more alternative routes than prior LLM-based planners while improving synthetic route quality and maintaining a high success rate.

A central goal of quantum machine learning is to identify applications in which small quantum circuits can improve task performance while remaining executable on real hardware. We investigate deepfake (synthetic) speech detection, a largely unexplored application in this area, where reliable confidence scores are as important as discrimination performance because deployment decisions are made at fixed thresholds and overconfident errors can be costly. We introduce QBFF, which routes a compressed speech representation through a small 8-qubit variational circuit (VQC) used as a drop-in bottleneck, with a matched classical bottleneck as the control. The quantum circuit yields better-calibrated confidence in most matched comparisons, including out-of-distribution evaluation, though the downstream deployment-cost benefit is model-dependent. We then run the same circuit on a real IBM quantum processor. To our knowledge, this is the first anti-spoofing pipeline to run on a QPU; hardware execution demonstrates feasibility, with lower subset EER and $C_{llr}$ but higher mean actDCF than the matched classical reference. Together, these results show that hardware-executable quantum bottlenecks can improve score reliability, with gains that vary by evaluation setting.

We study a system of two quadratically-coupled harmonic oscillators, prepared respectively in the ground state and in a coherent state. This toy model can describe for instance the dynamics of one mode of a quantum field evolving in a harmonically-oscillating classical background. In such situations, it is well-known that, depending on the background oscillation amplitude and coupling strength, the system may exhibit parametric resonance. Using the Feynman-Vernon influence functional, we analyze the interplay between parametric resonance and quantum entanglement by perturbatively computing the purity of the mixed state resulting from tracing over one of the oscillators. We find a runaway behavior of the quantum entanglement between the two oscillators, as well as a rapid breakdown of the validity of perturbation theory, in regions of parametric instability. This provides a quantitative confirmation of various statements in the literature and is highly relevant to questions pertaining to the validity of the semiclassical approximation in coupled quantum-classical systems.

This paper deals with the pullback measure attractor for a non-autonomous distribution-dependent stochastic lattice wave equation with nonlinear noise on $\mathbb{Z}$. We establish well-posedness in the energy space $\ell^{2}\times\ell^{2}$ and show that the solution generates a continuous non-autonomous dynamical system on $\mathcal{P}_{4}(\ell^{2}\times\ell^{2})$, which possesses a unique pullback measure attractor $\mathcal{A}$. Furthermore, we prove the convergence of the attractor $\mathcal{A}^{\varepsilon}$ as $\varepsilon\to 0^{+}$ to the attractor of the corresponding distribution-independent limit system.

Allreduce is a fundamental communication primitive in large-scale distributed training, and torus interconnects remain widely deployed in high-performance systems. We introduce SP-allreduce, a new allreduce algorithm for $d$-dimensional tori that exploits the symmetry of the topology and routes only along shortest paths. We call an allreduce algorithm fully bandwidth-optimal if it incurs no bandwidth overhead, no multihop-induced traffic inflation, and achieves perfect per-round load balance. SP-allreduce is based on a family of carefully symmetrized shortest-path trees that guarantees full bandwidth optimality in the $\alpha$–$\beta$ model. We prove that it uses the minimum possible number of rounds among all fully bandwidth-optimal allreduce algorithms on $d$-dimensional tori, and therefore the minimum running time in this class. In this sense, SP-allreduce refines Bucket allreduce: it preserves Bucket's key strengths while reducing the number of rounds by a factor of two. We further extend our approach to broader symmetric topologies and to communication trees of prescribed depths. Experimentally, the resulting algorithms improve over the best compared state-of-the-art algorithms for tori by up to $1.19\times$ on square tori and $1.65\times$ on rectangular tori across the tested configurations. On HyperX and spine-leaf networks, we demonstrate gains of up to $14.82\times$ and $1.72\times$, respectively, over the best evaluated baselines.

In this paper, we apply persistent homology to study finite topological spaces by associating persistence diagrams with their elements through the furtherness function. This approach enables us to classify elements according to their topological evolution within the space. We further establish a bottleneck stability theorem showing that the bottleneck distance between the persistence diagrams associated with two elements is bounded above by their mutual furtherness.

The first-order phase transition is central to electroweak baryogenesis and gravitational wave generation, yet thermal equilibrium effective potential analyses cannot describe the real-time non-equilibrium dynamics of the phase transition process. We use the Schwinger-Keldysh closed-time-path formalism of real-time finite-temperature field theory to study these effects. For the $\phi^6$ model, we derive the effective action and the corresponding Langevin equation. We further extend the formalism to the two scalar fields case, obtaining the coupled Langevin equations. These results provide a theoretical framework for the bubble nucleation simulation of the non-equilibrium phase transition process, laying the foundation for reliable predictions of baryogenesis and gravitational wave signals.

Road-network trajectory synthesis under local differential privacy (LDP) enables mobility data analysis without uploading raw trajectories to an untrusted server. Existing locally private trajectory synthesis methods mainly model whole trajectories, point sequences, or grid-level mobility patterns. However, they cannot handle non-uniform privacy risks caused by local road-network and trajectory contexts, such as endpoints, long-stay-associated regions, and low-out-degree branches. They also do not directly protect conditional next-hop choices on road networks, where a user's selected successor may reveal sensitive local movement intentions. To address these challenges, we propose TrajRACE, a locally private synthesis framework for road-network trajectories. TrajRACE uses start, segment-count, and transition reports to represent trajectories. For each transition, it keeps the current road segment as a public aggregation context and perturbs the true successor among legal successors on the client side. It then assigns risk-aware protection strengths to transitions according to local road-network and trajectory contexts. Finally, the server recovers population-level start, segment-count, and conditional transition statistics from heterogeneous privatized reports, and generates synthetic trajectories that follow the public road-network topology. Experiments on three road-network trajectory datasets demonstrate the effectiveness of TrajRACE in improving synthetic-data utility across distributional, query-based, and pattern-level metrics, while providing stronger protection for high-risk local transitions.

Elite athletics throwing provides a structured biomechanical action sequence together with an objective continuous outcome, namely throw distance, making it a suitable setting for performance-oriented sports analytics. In this work, we investigate whether explicit phase alignment of pose dynamics improves throw-distance prediction under realistic broadcast conditions. To support this study, we curate Throwing4, a dataset of 500 elite throwing clips spanning javelin, discus, hammer, and shot put, with corresponding two-dimensional pose sequences, frame-level phase annotations, and official distance labels. We formulate throw-distance prediction from broadcast-derived pose as a regression problem and compare full-sequence modeling against a phase-conditioned approach that partitions each attempt into four technique stages and fuses phase-specific motion representations. Experiments across multiple graph-based pose encoders show a consistent reduction in prediction error when phase structure is incorporated, indicating that phase-aligned modeling provides a more stable representation than treating the entire motion as an unstructured sequence. These findings suggest that explicit biomechanical phase information is a useful inductive prior for pose-based performance modeling in elite athletics throwing. Beyond score prediction, the dataset and protocol provide a basis for future work on fine-grained technique analysis, interpretable sports analytics, and robust modeling from broadcast video.

Behavior-conditioned animal vocalization generation aims to synthesize vocalizations from natural-language descriptions of animal behaviors, social interactions, and contexts. This task remains largely unexplored, with limited behavioral text supervision and substantial variation in vocalization duration. We construct the Behavior-Vocalization Dataset (BeVo) by integrating public bioacoustic data from six species with natural-language behavioral conditions derived from species-specific ethological and bioacoustic literature, totaling 31,191 events. We then adapt a rectified-flow-based generator with Event-Aware variable-length modeling, which encodes vocalizations at their actual durations and uses masks to prevent attention to padded latent positions and exclude these positions from the training loss. We evaluate vocalization discriminability, acoustic distribution fidelity, and language-aligned representation fidelity. Experiments show that task-specific adaptation improves all evaluated metrics compared with its zero-shot baseline, while Event-Aware modeling further increases vocalization-anchor macro accuracy from 59.86% to 61.99% and reduces Frechet Audio Distance from 1.40 to 1.22. Conditioning analyses further show that behavioral conditions improve target-anchor discriminability, while swapping behavior shifts generations toward the corresponding vocalization anchors. Demos are available on our project page.

This paper investigates the large deviation principle for 2D stochastic $g$-Navier–Stokes equations driven by multiplicative noise and defined on an unbounded channel-like domain. We first prove the existence and uniqueness of probabilistically strong solutions for the stochastic equation under small noise perturbations by means of the local monotonicity method. Subsequently, we show the well-posedness of solutions to the corresponding control equation and the convergence in distribution of the solutions for the stochastic equation as the noise intensity tends to zero. Finally, we establish the Freidlin-Wentzell type large deviation principle for the solutions of the stochastic equation by using the weak convergence method. To overcome the lack of compactness of Sobolev embedding on unbounded domains, we employ a scalar stream function together with a truncation technique to derive uniform tail estimates for the solutions of the control equation.

The fingering instability of thin-film flow is a significant obstacle to producing uniform and stable films in industrial processes. In the framework of lubrication theory, we investigate the linear stability of a gravity-driven thin film with an inclined contact line using numerical and asymptotic methods. The contact-line singularity is regularized by imposing the Greenspan slip model. As demonstrated by the results, although the contact-line inclination slightly enhances long-wavelength instability, it significantly suppresses the growth of spanwise finite-wavelength perturbations. Moreover, the most unstable wavenumber shifts toward the long-wave region as the inclination angle increases. The inclination of the contact line introduces complex eigenvalues, leading to traveling-wave modes, in contrast to the standing-wave modes observed for horizontal contact lines. We also performed asymptotic analysis in the limit of small wavenumbers, confirming the enhanced long-wave instability. The physical mechanisms underlying these stability characteristics are elucidated, revealing that the effects of contact-line inclination come from the eigenmode distortion and additional streamwise flux.

Randomized quantum algorithms for the estimation of expectation values of molecular observables offer a promising route to early fault-tolerant quantum advantage by trading circuit depth for an increased number of samples. However, this begs the question whether the resulting measurement overhead remains feasible for industrially relevant molecular systems. In this work, we benchmark the randomized sampling algorithm of Wang et al. [1] through numerical simulations on a hydrogen molecule and resource estimates for various active spaces of the cytochrome P450, an industry-relevant system in the pharmaceutical industry. Our numerical simulations reveal that the analytical error bounds significantly overestimate the actual sampling error, whereas the analytical bound on the required imaginary evolution time is comparatively tight, exceeding a direct DMRG estimate by roughly 44% for the active space where the comparison is feasible. The measurement overhead nevertheless remains prohibitively high: even for the smallest active spaces, the expectation-value estimation alone requires around 1010 samples, scaling to beyond 1015 for the largest active spaces considered. These results suggest that substantial improvements in measurement strategies or algorithmic scaling are necessary before randomized early-fault-tolerant algorithms become competitive for industrial quantum chemistry applications.

The KP-II equation was introduced by Kadomtsev and Petviashvili to explain the transverse stability of line solitary waves in shallow water. Using Darboux transformations, we prove the spectral stability of Y-shaped line solitons in an exponentially weighted space that captures the amount of waves receding behind the line solitons. We show that the spectrum of the linearized operator consists of the continuous spectrum of the zero background together with three residual spectral domains whose boundaries contain continuous spectra associated with the constituent line solitons. Exploiting the analytic dependence of the pairings with adjoint eigenfunctions on the spectral parameters, we prove that the orthogonality conditions associated with all adjoint eigenfunctions can be reduced to orthogonality conditions against three families of adjoint continuous eigenfunctions. Moreover, we investigate linear stability of Y-solitons in a subspace satisfying orthogonality conditions against adjoint continuous eigenfunctions associated with outgoing modulations along each line soliton constituting the Y-soliton.

This paper considers a four-dimensional Riemannian manifold with an additional circulant tensor structure $Q$. We prove that such a manifold is almost-Einstein if the curvature tensor preserves the action of the structure on the last two arguments. Additionally, we examine an associated Riemannian product manifold $(M, g, Q)$ where $P=Q^{2}$, deriving explicit expressions for the curvature tensor in terms of the metric and the product structure. The sectional and Ricci curvatures of these manifolds are determined using special bases generated by the images of a given vector under the structure $Q$. Finally, we construct a non-trivial example of a four-dimensional locally product manifold by embedding a four-dimensional hypersurface into a ten-dimensional Euclidean space.

Using the formalism of presheaf automata, we introduce new classes of automata for modelling concurrency that extend higher-dimensional automata and symmetric higher-dimensional automata. We study their languages and characterise these languages in terms of posetal cospans.

We propose Pathfinder-BP, a recursive decoder for topological quantum codes that combines belief propagation with controlled bit-flipping. An initial BP pass identifies and permanently fixes high-confidence errors. Upon convergence failure, the decoder recursively explores alternative error patterns by tentatively flipping low-confidence variables. On surface codes, Pathfinder-BP achieves a noise threshold and logical error rate comparable to minimum-weight perfect matching (MWPM). A crucial advantage of Pathfinder-BP is its highly parallelizable structure, which yields a worst-case decoding latency quasi-linear in the code distance, making it a suitable candidate for low-latency hardware implementation.

Nominal chip specifications do not capture the compute deliverable by heterogeneous distributed networks under communication, runtime and verification constraints. We formalize D-PIN Standardized Compute Units (D-PIN SCU) as a dynamic metering system for deliverable compute. The system combines five-dimensional resource normalization with topology-aware and runtime-corrected metering. A cluster-scale Roofline model yields a closed-form communication ridge, while runtime factors describe thermal, fault and load effects. We show that runtime deviations cannot be identified from nominal specifications alone. Uniform re-anchoring preserves relative metering results, and a simplex-constrained calibration algorithm estimates resource weights. The framework couples Proof-of-Physical-Work verification to a no-arbitrage condition for over-claiming and electricity-aware pricing. An end-to-end recomputation on the Zhangjiang Science Gate production node covers ten heterogeneous 8-GPU servers totaling 59.5 D-PIN SCU. Billed throughput of approximately 33,000 D-PIN SCU-hours is consistent with the physical upper bound at 78% utilization over 720 hours. Derived metrics are 3.02 kW per D-PIN SCU at the supply-and-distribution design boundary and 0.26 D-PIN SCU-hours per delivered kWh. Empirically calibrated chip values differ materially from datasheet projections; the published calibration is not independently reproducible from vendor specifications. Validation is limited to one production node, and the originality claim concerns the integrated system rather than the SCU terminology or individual components. The framework provides a basis for scheduling, verification and pricing of deliverable compute in distributed heterogeneous networks.

Smart meters are pivotal to the functioning of smart grids, as they measure and transmit energy consumption data that includes sensitive information about household energy usage patterns. While this data offers significant financial benefits, it can also reveal detailed insights into residents' daily routines, making the security of smart meters a critical concern. This paper introduces a novel remote attestation method for smart meters that ensures the integrity and privacy of data exchanged between smart meters and central servers while detecting unauthorized tampering. Our approach structures the communication data into a Merkle Tree-a hash data structure that facilitates efficient data integrity and authenticity verification. All operations involved in constructing the Merkle root and hashing critical data are executed in the secure environment of a Trusted Execution Environment (TEE) using OP-TEE, thereby maintaining the integrity and security of the smart meter's data. Comparing our experimental results with the related works highlights the superiority of our method in terms of resource efficiency, security robustness, and the feasibility of implementation on smart meters and resource-constrained devices, thereby enhancing the overall security and reliability of smart grid systems.

Although users may express different personalization needs across tasks, these task-specific behaviors can reflect shared underlying preference mechanisms. Existing LLM personalization methods often focus on surface-level interactions or undifferentiated histories, overlooking the shared mechanisms underlying these behaviors. We introduce PersonaFactors, a framework that distills reusable and interpretable personalization factors from heterogeneous user interactions. We construct a library of thirteen PersonaFactors by extracting evidence-grounded rationales and clustering recurring patterns across tasks, and develop a dynamic factor-guided generation framework that selects factors relevant to each user's history and current request. Across four personalization benchmarks and four backbone models, Dynamic PersonaFactor achieves the best score on every reported metric. Representation analyses further show that PersonaFactors and user history induce partially overlapping low-dimensional representations, and that their joint use yields complementary shifts. Taken together, these findings suggest that factor-based and history-based personalization can mutually reinforce each other.

Large language models are increasingly deployed in morally and legally sensitive contexts, yet little is known about how they internally represent distinctions between moral reasoning styles. We ask whether a model's differential response to rule-framed and consequence-framed moral questions reflects distinct internal structure or merely surface lexical sensitivity. We construct a 2$\times$2 factorial dataset crossing reasoning style (rule-based vs. consequence-based framing) with judgment polarity (permissive vs. prohibitive), drawn from three established moral reasoning benchmarks and evaluated on a cleanroom test set of novel phrasings that includes conflict pairs where the two framings predict opposite labels. Using linear probing and activation steering, we identify two orthogonal directions in the residual stream: Dir-A tracks reasoning style and Dir-B tracks judgment polarity, each near-blind to the other factor (cross-correlation $|r| < 0.04$). This double dissociation replicates across three instruction-tuned models from different developers at comparable proportional depth, and it is not knife-edged in layer: the directions are recoverable and causally effective throughout an extended band of the network. Dir-A can causally change judgments, and in conflicting scenarios, it selectively corrects errors in rule-based judgments without changing consequence-based judgments, suggesting that it does not simply make the model more likely to accept everything. Generalization beyond the training templates is partial: the direction transfers across benchmark sources, but its advantage over lexical baselines on novel phrasings is modest. We therefore frame the result as moral reasoning style sensitivity, a geometrically dissociable and causally effective structure, rather than internalized normative reasoning.

The polarization state of light provides direct access to ultrafast electronic, magnetic, and structural dynamics in materials. Transient wide-field microscopy with simultaneous amplitude, phase, and polarization sensitivity enables spatially resolved measurements of these dynamics, but its sensitivity is often limited by technical noise and background fluctuations. Here, we demonstrate transient holographic full-polarization microscopy with substantially reduced instrumental noise and sensitivity approaching the shot-noise limit. We compare holographic geometries using one, two, and four reference beams and find that the single-reference configuration provides the most favourable shot-noise-limited sensitivity for the derived polarization observables. Combined with two-camera detection, it simultaneously captures both orthogonal polarization components without distributing reference photons among multiple carriers. A self-referenced noise-reduction framework combining singular value decomposition and background subtraction further suppresses correlated noise. For spatially averaged regions, the baseline noise is reduced by factors of three-sixty, enabling detection of weaker transient responses at lower excitation densities and in weakly absorbing materials.