Papers

Let $n$ and $N$ be integers with $2\leq n<N$, and let $F:\mathbb{B}^n\to\mathbb{B}^N$ be a proper holomorphic map that admits a $C^{N-n+1}$-smooth extension to the boundary. We prove that if the CR Gauss map of its boundary restriction is generically degenerate, then $F$ is holomorphically equivalent, under composition with automorphisms of the source and target balls, to $V_m\oplus 0$ for some positive integer $m$, where $ V_m(z)=\big(\sqrt{\frac{m!}{\alpha!}} z^\alpha\big)_{|\alpha|=m} $ is the degree-$m$ Veronese map. This resolves a problem posed by Xiaojun Huang.

The IXPE mission enabled measurements of X-ray polarization in accreting binaries. The dependence of the X-ray polarization degree on the binary inclination and the optical depth of the emitting medium $\tau_0$ are the new important tools in X-ray Astronomy. We analyzed the black hole X-ray binary 4U~1630–47 as an example of one of the enigmatic sources in our Galaxy. We consider spectral and polarimetric observations of outburst episodes from 4U~1630–47 observed with the IXPE and the NICER. We carefully investigated a time interval of the IXPE observations during which 4U~1630–47 was in a single spectral state (the intermediate state) and found that the polarization degree did not change significantly during this time interval with the photon energy in the 2–8 keV range ($P\sim$ 6%). We interpreted this energy independence behavior in $P(E)$ as a result of the Comptonization (up-scattering) of soft photons in the flat Compton cloud. We argued that the X-ray source broadband energy spectra can be reproduced by a physical model composed of the Comptonized component and iron-line (Gaussian) component. Spectral analysis using this model shows that during the intermediate state the 4U~1630–47 spectrum was characterized by a photon index $\Gamma\sim 2.5$ and the electron temperature $kT_e\sim ~9 ~{\rm keV}$. In fact, this approach allows one to easily find the half-optical depth $\tau_0\sim 1.8$ of the Compton cloud in 4U~1630–47 applying the inclination $69\pm1^\circ$ and using the relation $(P,\cos(i),\tau_0)$ when solving the full radiative transfer equation for a flat Compton cloud by an iterative method. This approach is also accompanied by the fact that the polarization degree is independent of the photon energy, which we demonstrate observationally, at least for 4U~1630–47 in the 2–8 keV range.

Implicit Large Eddy Simulation (ILES) is a well established approach for modeling high-Reynolds-number turbulent flows by numerical solution of the inviscid Euler equations in 3D. These turbulent flows cover a wide range of Mach numbers, including within individual simulations. For example, Richtmyer-Meshkov instability simulations transition from a high-Mach compressible regime to a low-Mach quasi-incompressible regime during the late-time evolution. These flows motivate the development of a semi-implicit all-Mach numerical method, which is subject to a CFL stability constraint corresponding to the fluid velocity, but independent of the speed of sound, and can be applied efficiently and accurately across the whole spectrum of Mach numbers. Furthermore, finite-volume discretization is chosen to maintain strict conservation and ensure robust and accurate capturing of strong shocks. The method is validated in both high- and low-Mach regimes. Validation is also presented for stratified atmospheres in the presence of gravity, and the method is thus shown to be applicable to buoyancy driven flows. We present favorable comparison against previously published ILES simulations of multimode Richtmyer-Meshkov instability from the $\Theta$-group collaboration, Thornber et al., (arXiv:1706.09991).

Pre-generation success probes estimate response correctness from a language model's hidden activations before decoding, enabling cost-aware routing. While prior work has demonstrated their utility primarily on English inputs, we study their reliability across languages along three dimensions: (1) whether they preserve the ranking of likely successes and failures (DISCRIMINATION); (2) whether they retain probabilities that match observed success frequencies (CALIBRATION); and (3) whether they produce scores comparable enough across candidate models for cost-aware multilingual routing (UTILITY). Using 3,000 MATH problems in 10 languages and 8 open-weight model configurations, we compare cross-lingual transfer from English-trained probes and equal-budget pooled multilingual probes. English-trained probes retain useful cross-lingual discrimination but become less well calibrated after transfer. Pooled multilingual supervision improves both properties and yields more reliable estimates of success. In routing experiments, the pooled router achieves a 0.7% higher test success rate while reducing modeled cost by 13.0% relative to always selecting the model with the highest average success. These results show that multilingual routing requires success estimates that remain well calibrated and comparable across languages and models.

Micro unmanned aerial vehicles (micro-UAVs) are small enough to reach confined spaces that larger robots cannot access, but too small to carry the sensing and computing power required for autonomous flight. We move the localization stack entirely off the aerial platform onto a quadruped robot with a 7-degree-of-freedom (DOF) arm, which supplies the micro-UAV (27 g bare, 42 g with fiducial markers) its full 6-DOF pose. A camera at the arm's end-effector detects AprilTag fiducial markers on the drone and composes that observation with the quadruped's own self-localization to place the drone in a shared map frame, so the ground robot localizes its partner, rather than only tracking it relative to the camera. The arm acts as an actively-controlled observer, repositioning to keep the drone in view as both robots move; the drone carries only an inertial measurement unit and fuses the external pose to fly commanded setpoints. In lab flights the external pose is accurate to 12-16 mm, enough to fly the drone autonomously within 2-5 cm of motion-capture-fed control. Having the quadruped actively follow the drone reduces the tracking error from 11.0 cm to 6.9 cm by holding the camera in the close range, where the markers are most accurate.

We study the fundamental solution operators associated to Kolmogorov equations with symmetric non-local diffusion with rough kernels. We prove that these operators are induced by measurable kernels and prove sharp upper bounds in terms of a kinetic profile governed by the interplay of transport and non-local diffusion. Our proof combines the small-jump with loss decomposition of P.-A. Meyer with the ideas of Nash. The proof is purely analytic.

We study the sampling allocation of LinUCB in the small-gap regime, where the reward gaps are of order at most $n^{-1/2}$ over the decision horizon $n$. This scaling captures the hard instances underlying worst-case regret lower bounds, for which LinUCB is known to be near optimal up to logarithmic factors in $n$. Using a mean-field perspective, we characterize this allocation through the empirical sampling distribution, a macroscopic object that averages the effect of adaptive decisions over the horizon, and identify its limit as $n\to\infty$. We establish that in this regime, the empirical sampling distribution induced by LinUCB converges to the set of D-optimal designs. This central result reveals that, in the small-gap regime, LinUCB not only achieves near optimal minimax regret but also allocates samples in a way that is asymptotically efficient for learning the reward parameter, thereby connecting regret-driven online learning with information-efficient experimental design. Building on the optimal design limit, we obtain two useful consequences. First, we refine the asymptotic regret analysis of LinUCB in the small-gap regime by characterizing its leading-order constant in the limit. Second, we show that, despite LinUCB's adaptive sampling strategy, the regularized least-squares estimator satisfies a central-limit-type theorem in the small-gap regime, thereby enabling valid statistical inference for the reward parameter.

Online agents can improve future service by constructing reusable tools, guidance, or model states, but this work competes with current requests for the same GPUs. Exploiting idle compute for self-evolution faces a fundamental systems constraint: benefits arrive only after an artifact is published and used, while pausing evolution leaves service capacity waiting for memory release and runtime recovery. An investment worth completing may therefore be worth postponing. We present LearnSched, a state-aware scheduler that brings the reuse window of a capability and the timely return of compute into a common investment model. LearnSched incorporates candidate progress, checkpoint overhead, and the current recovery path into action values. Under the same information and capacity constraints, it uses one-step counterfactual rollout to choose progress, checkpointing, or waiting relative to a fully costed window policy. We characterize handoff costs through independent A100 component measurements and evaluate action selection in 1920 paired finite-model scenarios. When cold recovery is expensive, waiting for longer execution windows can improve net value by avoiding handoffs; in evaluated warm-recovery scenarios, the strong window policy already achieves the same value. Retaining recoverable state can shorten capacity return and reduce the need for complex evolution scheduling.

Learned models for industrial control are usually judged by aggregate accuracy, but accuracy at the component level does not guarantee safety once it is embedded in the system it is meant to serve. We study this gap on a behavior-cloning task: imitating an expert Nonlinear Model Predictive Control (NMPC) policy for load-following of a Pressurized Water Reactor (PWR), an industrial system with tight safety constraints. We propose a structured architecture encoding variables from each timescale into separate latent spaces, reflecting the physical decomposition of the system, before training a controller to imitate the expert on the product latent space. On long-horizon rollouts, separated embeddings improve both accuracy and feasibility compared with a shared-embedding baseline. Sensitivity analysis further shows that our model yields interpretable representations aligned with the system's physics. However, standalone deployment still leaves several percent of trajectories infeasible regardless of the architecture. Using our method to warmstart the NMPC optimizer rather than acting standalone, we recover full feasibility and near-optimal cost while still cutting computation time by $\sim$15% relative to the expert controller, and even more for abrupt operating changes.

Autoregressive generation is natural for language, where predicted tokens can be directly reused as the next prediction state, but trajectory forecasting lacks such a clean token: motion is continuous, multimodal, and expressed in local coordinate frames that evolve with the predicted trajectory. We present MoCAR (Motion-code Coordinate-aware AutoRegression), a decoder-only framework that casts trajectory forecasting as next-code prediction in a coordinate-aware continuous latent space. MoCAR learns a continuous motion-code space from endpoint-normalized trajectory segments, where each code jointly captures local trajectory geometry and the reference-frame transition induced by that segment. Historical motion codes are used as a teacher-forced prefix, future codes are generated autoregressively under temporal, map, agent, and mode interactions, and predicted codes persist in latent memory while decoded endpoints update the local scene context. This enables rollout without trajectory-space re-tokenization, trajectory queries, goal candidates, or proposal-and-refinement pipelines. On Argoverse (AV) benchmarks, MoCAR achieves top-tier performance with a simple single-stage architecture, transfers strongly from AV2 to AV1 in zero-shot evaluation, and improves on turn-heavy scenarios. Ablations confirm that the learned continuous motion-code space, latent alignment, weak KL regularization, and joint tokenizer-predictor optimization are essential for stable latent autoregression.

Trapped particles in hollow-core fibers enable long-range sensing, though advanced control of their in-fiber motion, such as multimodal cooling and squeezing, remains challenging. Here, we present optical interference-based adaptive techniques for controlling the motion of a fringe-trapped silica nanoparticle inside a fiber. After feedback-cooling its axial and radial motion, we induce axial delocalization (position anti-squeezing) via two complementary approaches. First, non-adiabatic fringe suppression expands the position variance by 11.83 ($\pm$0.7) dB to that of the initial cold-state while retaining Gaussian statistics. Second, multi-pass particle positioning at dark fringes increases the delocalization to 13.23 ($\pm$0.5) dB relative to the cold-state's variance via dark inverted optical potentials, in agreement with our Wiener stochastic model. Stronger delocalization produces non-Gaussian states of motion. Unlike prior inverted-trap implementations, our method requires neither auxiliary optical traps nor charged particles in Paul traps. Our results demonstrate fringe-trapped particles in hollow-core fibers as a versatile platform for long-range sensing and macroscopic quantum physics.

In this work, we propose and analyze low-order nonconforming face-based finite element methods for elliptic problems on complex hybrid three-dimensional meshes, composed of arbitrary polytopal cells. The focus is on scalar problems, but the approach can be extended in a tensor manner to design generalizations of the classical Crouzeix–Raviart and Rannacher–Turek methods for mixed formulations. We adopt a nonparametric construction of the local approximation spaces. We demonstrate that it avoids the degradation of approximation properties commonly induced by parametric mappings on distorted meshes. Using these approximation spaces, we define finite element schemes for very general cells. We further demonstrate that, on strongly degraded meshes, an approach inspired by discontinuous Galerkin methods enables removing consistency errors, even in the presence of potentially curved faces. Explicit examples of the construction of the polynomial space are given for prismatic (of triangular basis) and pyramidal (of quadrilateral basis) cells. The resulting schemes are shown theoretically and numerically to be stable and convergent, with an optimal convergence rate.

Large Language Models (LLMs), Retrieval-Augmented Generation (RAG), and AI agents are increasingly embedded in operational business processes. Yet Business Process Management (BPM) curricula and frameworks still largely treat artificial intelligence (AI) as an add-on technology, leaving graduates (as potential future process developers) unprepared to reason about AI as a first-class design element of end-to-end processes. This paper addresses that gap by proposing AI-decision checkpoints: explicit moments in a process development trajectory where process developers identify AI-candidate sub-processes, assess expected effects on time, cost, quality, and flexibility, consider legal and organisational constraints, and document a reasoned decision to adopt, constrain, or reject specific AI components. The checkpoints are instantiated through a fictitious customer onboarding process as a BPM Teaching Case, embedded in a lifecycle-driven framework spanning six modules that combine process modeling, simulation, workflow execution with AI agents, and process mining, with each module's output serving as the next module's input. A preliminary formative reflection draws on instructor observations, submitted artifacts, and discovered process maps from learning-management-system logs. These exploratory observations suggest that the approach supported clearer distinctions between task-level automation and process-level value.

We study the $\Gamma$-convergence of a general class of anisotropic phase-field functionals modelling cohesive fracture in the multidimensional, vector-valued, and geometrically nonlinear setting. The limiting energy comprises a quasiconvexified bulk term, a Cantor part governed by its recession function, and a cohesive surface energy density defined via an asymptotic cell problem. After establishing the structural properties and equivalent formulations of the limiting surface density, we show that, whenever the recession function of the elastic energy is given by the square of an operator norm compatible with the phase-field anisotropy, the vectorial cell problem explicitly reduces to a one-dimensional variational problem. As a consequence, we extend the cohesive law reconstruction procedure of \cite{alessi2025phasefieldpart2} to the vectorial framework, explicitly designing phase-field potentials that recover prescribed classes of anisotropic, jump-dependent traction-separation laws of the form $g(\zeta,\nu)=\gamma(\nu)g_{\scal}(|\zeta|)$.

Conversational sales agents must distinguish questions about products from purchase commitments, preserve explicit requirements, and ground the next action in current business information. We report an aggregate census of 725,219 records in an accessible conversation table provided by Aiso and develop a reference architecture for this setting. All records have distinct non-null conversation hashes. Existing metadata labels identify 41,800 commercial records (5.76%) and 2,387 transactional records (0.33%); their union contains 44,187 records (6.09%). Within the commercial category, 54.31% are labeled English, 41.59% have recorded depth of at least two, and 13.51% have depth of at least four. Commercial-label prevalence varies from 4.87% to 6.35% across three source batches. These measurements motivate explicit separation of corpus inventory, commercial relevance, training eligibility, and observed business outcomes. The proposed architecture combines business-grounded knowledge, provenance-bearing conversation state, and a constrained next-action policy. A quality specification addresses source rights, privacy, label validation, deduplication, and training-test separation. The study is a metadata audit and technical design, not a validation of label accuracy, model training volume, or sales conversion. No raw conversation text or personal identifiers are released.

LLM agents are starting to own the full customer experience. Soon, LLMs may be selling and buying on behalf of companies and customers respectively. Small models are more cost-efficient at scale, but can reinforcement learning train them into competent sellers? We train four Gemma 4 checkpoints (2.3B to 31B effective parameters) with GRPO on a programmatic utility reward for bilateral multi-issue bargaining, and evaluate every arm on the same 1,152 negotiations against two frontier buyers it never saw in training. With the same learning rate ($10^{-6}$) for every size, the gain of the RL model over its base rises from $+0.001$ at 2.3B to $+0.078$ at 31B. Each size was trained once and the two smallest checkpoints use a different architecture, so we fit no scaling law. Tripling the learning rate, with the same or fewer training steps, improves on the shared rate at every size by $+0.032$ (2.3B) to $+0.081$ (4.5B). In exploratory comparisons with two frontier models run as sellers, the 12B seller trained at the tripled rate scores above both, though its untrained base already scores as high as they do. The 4.5B seller at that rate shows no detectable difference from either and fits on one 48 GB GPU. A further 2.3B arm at ten times the shared rate raises pooled score, but its gain concentrates on the evaluation buyer that shares a model family with the training pool. These results suggest tuning the learning rate before concluding that a small model cannot learn to negotiate, and testing against buyers from more than one model family.

The manipulation of non-Abelian anyons is a foundational ingredient for topological quantum computing and underlies topologically protected quantum memories. Here we analyze a quantum-impurity-based protocol for braiding non-Abelian anyons. Exploiting mobile impurities with addressable internal spin states, we provide a concrete protocol to experimentally measure the matrix elements of the $R$ and $F$ matrices that characterize the non-Abelian statistics of anyons. In particular, focusing on the case of Ising and Fibonacci anyons, we illustrate how the anyonic nature of the excitations emerges directly in the Ramsey interferometer we propose. Building on this framework, we further introduce a conventional-to-topological qubit encoder to transfer quantum information from conventional flying qubits into a topologically protected quantum memory.

We develop a finite-temperature field-theoretical description of the Casimir interaction between two perfectly conducting boundaries with the intervening gap filled by a two-component quantum Boltzmann plasma embedded in a homogeneous dielectric background. The quantum motion of the charged particles is represented by imaginary-time worldlines, while the electromagnetic field is treated in the Gaussian approximation. The resulting partition function contains distinct longitudinal and transverse response contributions and an additional quantum-induced source term with no classical counterpart. In the classical limit the theory reduces to the usual Casimir-screening picture: the static longitudinal mode is Debye screened, whereas the transverse electromagnetic sector is unaffected by classical mobile charges. Finite thermal de Broglie wavelengths modify the nonzero Matsubara response and produce the leading quantum correction to the force. In the static long-wavelength limit, the longitudinal dielectric function has, through the leading quantum correction, the same expansion structure as the Lindhard response of a degenerate electron gas under the correspondence $k_{TF}\leftrightarrow\kappa_D$ and $k_F\leftrightarrow\lambda_{dB}^{-1}$, while the static transverse response reproduces the leading high-temperature Landau orbital diamagnetic susceptibility.

In this paper, we study finite-mode Gaussian quantum Markov semigroups with a faithful invariant Gaussian state. We present an algebraic characterization of the complete modified logarithmic Sobolev inequality relative to the fixed-point algebra, and we obtain the corresponding optimal constant under the assumption that the drift evolution $e^{t\mathbf{Z}}$ converges as $t\to\infty$. We establish hypercontractivity for such quantum Markov semigroups with Hurwitz drift, in the sense that every fixed pair $1<p<q<\infty$ is attained at sufficiently large times. We find that the usual reverse hypercontractive curves with a positive rate from time zero can fail in general, but every fixed pair $1/2 < p < q < 1$ is attained at sufficiently large times under the Hurwitz assumption. We also systematically investigate the $p$-logarithmic Sobolev inequality and show that the $p$-logarithmic Sobolev constant can be negative for $0 < p < 1/2$.

In this article, we show that the characterization of reversible rings and various other rings related to them such as i-reversible rings, semi-commutative rings and i-domains in terms of properties related to idempotent elements known in the literature can be generalized to exactly the same properties for quasi-idempotent elements.

Foreground cleaning is a major challenge for H\,i 21 cm intensity mapping, particularly in the presence of a chromatic instrumental beam that couples angular foreground structure into the frequency direction and under incomplete sky coverage. We introduce JESSNet, a beam-aware, mask-compatible foreground-cleaning framework for single-dish intensity mapping. Building on SDecGMCA, JESSNet introduces three main extensions: a multiscale angular decomposition with scale-dependent mixing matrices and channel selection, a learned spherical-wavelet ”learnlet” sparse regularization operator, and a mask-constrained reconstruction applicable to Galactic masks and survey footprints. We test JESSNet on simulated SKA-MID-like observations over $900\text{-}1300\,{\rm MHz}$ ($z\simeq0.09\text{-}0.58$), including H\,i emission, Galactic and extragalactic foregrounds, thermal noise, and an oscillating MeerKAT-inspired chromatic beam. The multiscale reconstruction accurately recovers the angular and frequency power spectra of the input H\,i signal, improves the recovery at low and intermediate angular scales relative to a single-scale implementation, and reduces residual beam-induced spectral features. JESSNet also remains effective for a synthetic survey footprint, preserving the H\,i power spectra under incomplete sky coverage. The code and analysis pipeline are publicly released to facilitate reproducibility and future applications to simulated and observational data.

Cutting is a recurring operation in robotic au tomation, from food preparation to surgical tissue resection. For highly compliant targets, blade motion may deform or displace the material rather than advance the intended cut, while progressive failure changes load transfer and subsequent tool tissue interaction. A computational description must therefore connect cut formation to the evolving mechanical response, rather than specifying the incision independently of material failure. We present STC-MPM (Soft Tissue Cutting with the Material Point Method), a framework that couples finite-strain deformation, history-driven continuum damage, and configurable post-failure treatment. Damage initiates only when a tensile strain history exceeds a material threshold within the blade process zone. Progressive degradation changes stress transmission before the post-failure policy is applied, while the remaining tissue continues to deform, move, and interact with the tool. Numerical scalpel studies using delayed particle deactivation follow this response through insertion and withdrawal. Under identical blade motion, reducing the damage-rate limit leaves the first recorded damage time unchanged but delays and increases peak reaction force, delays particle deactivation, and increases the fixed-cohort displacement statistic at maximum insertion. The matched comparison illustrates how post-initiation failure evolution alters subsequent tool loading and recorded material motion. STC-MPM thus supports joint analysis of cut formation and accompanying tissue response without a pre-inserted cutting interface.

We give quantum algorithms that estimate the log partition function of an $n$-qubit quantum Hamiltonian to additive error $\epsilon$ with Heisenberg scaling, while minimizing calls to thermal-state purification preparation. The input is a block encoding of $H$, unitaries preparing thermofield-double states at chosen inverse temperatures, with their inverses and controlled versions, and an adaptive slowly-varying cooling schedule. The system-size bounds below assume $0\preceq H\preceq hI$ with $h=O(n)$ and block-encoding normalization $\alpha=\Theta(h)$. For constant inverse temperature, we show that such a schedule of length $O(\sqrt n)$ always exists and can be generated from overlap estimates with $\widetilde O(\sqrt n)$ further preparations. Given such a schedule and classically chosen temperatures, we estimate the log partition function with $\widetilde O(n^{5/4}/\epsilon)$ expected calls to state preparation and to the block encoding, improving by $n^{1/4}$ on non-adaptive strategies. To achieve our result, we use a recursive-doubling identity that expresses each schedule increment as a weighted sum of logarithms of thermofield-double overlaps and of short imaginary-time steps implemented by quantum singular value transformation combined with previously developed variance reduction techniques. If thermofield doubles can be prepared controlled on a temperature register, a quantum inverse-binomial log estimator combined with quantum mean estimation reduces both query counts to $\widetilde O(n/\epsilon)$. We show that the latter scaling is optimal in block-encoding queries up to polylogarithmic factors and discuss end-to-end complexities for one-dimensional partition-function estimation.

Remote sensing images often carry composite degradations, in which haze, cloud, noise, blur, low light, and low resolution coexist. Restoring them requires deciding which tool to apply, in what order, and when to stop, yet no clean reference is available at inference time to verify these decisions. All-in-one models trained on single degradations converge to a narrow PSNR band as degradations accumulate. To formulate real-world remote sensing restoration as a traceable trajectory, we present EORestore-Agent, which replaces this unmeasurable objective with reference-free, verifiable per-step decisions. A fine-tuned vision-language model reports all residual degradation types, whose tool pools are scored together, so the restoration order emerges from step-wise selection. A relative quality scorer, trained with full-reference supervision on synthetic degradation chains, predicts the changes in PSNR, SSIM, and LPIPS from the current image to each candidate. A step is accepted only when no predicted change is negative and the predicted PSNR gain is positive. Otherwise, the agent keeps the current image. On a synthetic Landsat-8 benchmark with six degradation types, EORestore-Agent improves PSNR by 2.3 to 3.2 dB over the strongest retrained all-in-one baseline on composites of two to six degradations, whereas zero-shot natural-image agents fall below the degraded input in PSNR in 17 of 18 settings. Replacing the learned scorer with no-reference quality differences costs 1.1 to 4.6 dB. The remaining harmful steps are small and cluster near the acceptance threshold. Sentinel-2 examples illustrate transfer to real atmospheric degradation without retraining.

Learned feature representations in neural networks often contain structure beyond that directly used by the final task output. We study this structure using $\textit{passive probes}$ that apply fixed, untrained, property-independent projections to representations as they evolve during training. We motivate this approach through the task of prediction on $S^2$ where equivalent vector and Hermitian parameterizations reveal an additional loss-invariant trace coordinate. This motivates a general construction in which fixed random projections serve as observers of learned features. Because the observer is loss-invariant and independent of the property being studied, changes in accessibility reflect changes in the representation relative to the fixed observer rather than adaptation of the observer itself. We show that ensembles of passive probes can directly reflect task-relevant information such as target alignment. Under our constructions, the accessibility of eventual difficulty evolves differently across tasks. It increases during training in the regression tasks of surface-normal estimation and image inpainting but remains near its initial level in image classification. Comparisons with learned linear probes further show that recoverability and passive accessibility can evolve differently during training. Together, these results show how passive probes can separately characterize changes in representation geometry and the accessibility of eventual task difficulty.

Reliable verification of squeezed-light sources is essential for photonic quantum technologies. High-dimensional quantum optical networks demand single-mode, pure squeezed states which, in principle, can be generated through careful engineering of parametric down-conversion-based (PDC) sources. However, due to the imperfections of optical components, the generated state is never pure and commonly contains internal mode structure. Existing characterisation techniques, based on either joint spectral intensity or normalized second-order correlation functions, work under the assumption that the probed state is pure, and therefore can result in incomplete or misleading assessment of the source quality. Here, we develop a device-agnostic characterisation framework for PDC-based sources with distinguishable signal and idler fields. Treating the source as a black box producing a mixed multimode Gaussian state, we find a lower bound for the state purity, which can be experimentally determined via mode-unresolved measurements. Our approach enables a holistic assessment of realistic squeezed-light source quality without requiring a detailed model of the nonlinear device and its surrounding optical network, providing practical benchmarks for optimizing sources for state-of-the-art quantum applications.

Autonomous coding agents now resolve a substantial share of real-world GitHub issues. However, passing functional tests differs fundamentally from producing a high-quality contribution acceptable for merging. Mature open-source projects publish repository-specific contribution policies, spanning style, git, testing workflows, to ensure code quality and long-term maintainability. Because existing benchmarks evaluate patches solely on unit tests, agent compliance with repository governance remains unknown. In this paper, we introduce SWE-CC, a benchmark evaluating code and process compliance in autonomous software engineering. We develop a semi-automated pipeline that converts developer documentation across 12 open-source repositories into 823 machine-checkable atomic policies. SWE-CC introduces two features: 1) lightweight, deterministic checker functions that represent each policy, 2) a comprehensive auditing mechanism that inspects both agent runtime behaviors and final deliverables. We evaluate the compliance of agent workflows in 500 end-to-end software contribution tasks extended from SWE-bench Verified. Our evaluation of four LLMs under two agent scaffolds shows that modern agents suffer from coding compliance issues: although agents produce functionally correct patches, they still violate 43.1 percent of applicable project policies, with nearly half of all violations occurring during intermediate execution steps. These results show that functional correctness does not guarantee real-world readiness, highlighting that future software engineering agents must reliably conform to repository governance to enable safe and trustworthy deployment.

Multi-agent systems built on large language models (LLMs) restate observations as a matter of course: relays forward them, shared boards repeat them and discussion rounds echo them. An aggregator that pools such messages should count sources, not statements. We convert a reported probability into units of independent readings, which assigns every restatement a copy weight, 0 for an aggregator that counts sources and 1 for one that counts every statement, and yields the implied decision under any cost structure. Three testbeds hold the evidence fixed and vary how it is restated: message logs with an exact Bayesian oracle, web documents with appended copies, and logs written by LLM agent teams under four communication protocols. Across four models from three providers, a forwarded copy counts for 0.06 to 0.42 of a new reading, mostly because some replies count every statement. On 5% to 40% of logs that state one reading three times, the reported belief implies an early commitment that the oracle never makes. The models that count copies least and most on controlled logs do so on web copies and agent-written logs as well. A one-paragraph declaration of what a copy contributes brings the copy weight on controlled logs to 0.08 or less. A rule that has agents refer to readings instead of restating them cuts belief-implied early commitment from 11.2% to 1.1% and preserves genuine corroboration.

An agent whose tool keeps returning nothing useful should stop relying on it. In a retrieval environment with controlled source failures, we separate how agents judge results from what they do. We compare stopping at the same step after longer and shorter runs of results the agent judged useless; this contrast is zero for clock- or deadline-driven stopping. Where we record their judgments, the seven agents we test call a failing source's results useless 97-100% of the time, yet most of them rarely stop on that judgment. Prompt cues change when they stop but not what they stop on. Permission to answer from memory and a reasoning mode can bring early stops regardless of evidence, a stated budget moves the 7-8B models' stops to the deadline, and a stopping rule or call cost in the prompt is followed at most partly. Stopping follows the evidence only when the harness enforces an integration step that makes the agent answer after five consecutive results it judged useless. This step raises failing-source success for every model, keeps the stopping point fixed when the budget doubles, and needs no extra judgment call when the agent states its judgments. A pre-registered replication on 300 fresh questions confirms the dissociation and the rule's effect.

Large language models (LLMs) can reason over scientific literature to devise design strategies, yet fail to reliably implement them for biological sequences. While protein generative models learn sequence patterns, they lack the capacity to incorporate literature evidence for multi-step reflective reasoning, forming an evidence-to-execution gap between scientific reasoning and sequence manipulation. We present EASER (Evidence-Aware Sequence Engineering with Reflection), a reflective agent bridging reasoning and sequence generation via a learned property interface of offline-trained, fixed low-rank matrices. The agent steers a diffusion generator by combining these matrices, proposing intervention hypotheses (anchors, editable positions, control coefficients) grounded in retrieved evidence, sequence context and past results. A Probe-and-Steer mechanism validates interventions and allocates samples according to predicted property responses, with outcome reflection informing subsequent decisions. Evaluated on multi-objective antimicrobial peptide design (optimizing activity, non-hemolysis and non-toxicity), explicit hypothesis formulation delivers better multi-objective performance than direct action generation under identical decision conditions. Ablation studies verify the importance of evidence retrieval, episodic history, reflection and Probe-and-Steer. Over six repeated trials, EASER obtains the highest mean hypervolume and lowest mean IGD+ on screened candidates compared with competing baselines. Our work demonstrates how an executable property interface and iterative feedback link scientific reasoning to targeted peptide sequence generation.