Papers

Hidden-heavy pentaquarks can be identified as bound states in diabatic Born-Oppenheimer potentials that are repulsive color-Coulomb potentials at short distance, cross below a heavy-hadron-pair threshold, and then approach that threshold at long distance. We identify the pentaquarks $P_{c\bar{c}}(4312)^+$, $P_{c\bar{c}}(4337)^+$, $P_{c\bar{c}}(4440)^+$, and $P_{c\bar{c}}(4457)^+$ as the ground-state quartet for $c \bar c$ in the two coupled potentials associated with a spin-$3/2$ adjoint baryon. We predict that their $J^P$ quantum numbers are $\frac{1}{2}^-$, $\frac{3}{2}^-$, $\frac{3}{2}^-$, and $\frac{5}{2}^-$, respectively. Our prediction for $P_{c\bar{c}}(4457)^+$ differs from almost all previous predictions. Our prediction that there are only four hidden-charm pentaquarks differs qualitatively from that of simple molecular models that there should be seven. We identify the strange hidden-charm pentaquarks $P_{c\bar{c}s}(4338)^0$ and $P_{c\bar{c}s}(4459)^0$ as two members of the ground-state triplet for $c \bar c$ in the potential associated with a spin-$1/2$ strange adjoint baryon. We also present model-independent predictions for pentaquark decays into charmonium plus a light baryon and into a charm baryon and a charm meson.

Ultrasound is widely used, but acquisition quality is heavily dependent on the operator's knowledge and expertise. With demand for examinations outpacing the supply of trained sonographers, operator-guidance systems aim to close this gap by instructing a less trained user how to move the probe toward a target view. In this paper, we propose a retrieval-induced latent transition model for ultrasound acquisition dynamics, formulating ultrasound operator guidance as multi-step planning and retrieval in a world model. Using a V-JEPA 2.1 backbone, observations are first encoded into a latent space where anatomically related views lie close together. We then retrieve similar views from a reference database containing encoded latent states and corresponding probe positions and orientations. Rather than learning a parametric transition function, we directly use physically executed transitions from the database to establish our nonparametric, retrieval-induced transition model that supports receding-horizon planning. At deployment, guidance is generated from the live ultrasound image feed alone, without any probe tracking hardware. Applied to carotid ultrasound, the proposed planner reaches the target view in 86% of retrospective closed-loop episodes, versus 52% and 43% for representative baselines, outperforming both on every target view, including the challenging longitudinal internal and external carotid artery views. A prospective feasibility study on unseen volunteers, run in real time on a CPU using distillation, reaches 83% target-view reachability. Because planning is driven by proximity to any encodable goal latent, the same world model can navigate back to any previously acquired, patient-specific frame, supporting reproducible longitudinal imaging for e.g. perioperative or follow-up monitoring.

Two coefficients enter nearly every quantitative argument about monetary policy: the Phillips-curve slope and the policy rule's response to inflation. For seventy years, each generation that estimated them identified the coefficient conditional on assumptions it maintained, and its successor relaxed one of them, naming the bias this exposed. This survey reconstructs the two programs as a single identification problem. It distinguishes five senses in which a coefficient has been called identified, reads each generation of the curve and of the rule in that key, and shows that the programs fail together because each borrows its identifying condition from the other. The same sequence ran on the latent inflation target and on the optimal rule. The survey closes on what the credible designs license when they leave a range rather than a point: bounds, signs and orderings, and decisions designed for a range.

We study contextual queueing bandits, where a learner schedules jobs while learning unknown service rates modeled by logistic functions of job-server features. Performance is measured by queue length regret, the expected excess queue length at round $t$ relative to an oracle that knows the service rates. Existing decaying-regret guarantees either have a suboptimal decay rate or require a known fixed horizon. They also assume context-wise slack and a strictly positive minimum eigenvalue of the feature covariance. In this paper, we propose WISE (Widest Interval Selection with Elimination), achieving rate-optimal $\widetilde{\mathcal O}(t^{-1/2})$ queue length regret at every sufficiently large time without knowing the horizon. We assume capacity slack, meaning that expected incoming workload under best-server service is below service capacity, and impose no covariance lower bound. Our analysis uses a workload potential measuring the expected service attempts needed by waiting jobs on their best servers. Its drift on nonempty rounds combines a negative term ensured by capacity slack with errors from suboptimal service choices. Then an elliptical potential count bounds how often WISE selects wide confidence intervals, thereby limiting the number of rounds with large service errors. We also sharpen the arrival-rate dependence of an existing lower bound and make its dependence on feature dimension and server count explicit. We prove another lower bound that quantifies the increase in regret as the normalized capacity slack decreases; to our knowledge, this is the first such lower bound for CQB. Simulations show small regret even when context-wise slack fails.

We study transverse polynomials along polar arcs of reduced holomorphic plane function germs. We prove that, under bi-Lipschitz right equivalence, matched tangential polar arcs have the same transverse order and their transverse polynomials agree up to nonzero rescalings of the source and target, at every rational scale $1<q<d(\gamma)$, where $d(\gamma)$ is the gradient canyon degree. This range is sharp as the result can fail at both endpoints, even under analytic right equivalence. As an application, we recover the bi-Lipschitz invariance of the augmented Newton polygon proved by Migus–Păunescu–Tibăr. We also give an example showing that the transverse polynomial contains new information not detected by the polar-value invariants and augmented Newton polygons.

Small independent (indie) game startups face fierce competition when developing games with limited time, budget, and resources. Continuously testing and experimenting with new ideas, features, and content is essential to validate and guide development toward a viable and successful game. Such continuous experimentation (CE) requires access to users and user data, which is often very limited in the early stages of development, both in general and, in particular for indie game companies. We provide insights and recommendations on how to perform experimentation and game testing in pre-release stages of indie game development based on an exploratory interview survey with 10 independent game developers from different companies. We provide a framework for CE that highlights key parts to consider when planning and implementing an experiment. Current practices for experimentation within indie game development are also described for the early development stages of ideation, preproduction, feature production, and content production. Our empirical findings offer insights and recommendations into how to conduct experiments during pre-release stages of game development, specifically for Indie game companies. Other software startups, particularly those with a value proposition centred around interactive and satisfying user experiences, can also benefit from our research.

Formally reasoning about Rust programs requires a rigorous formal semantics, especially in the context of deductive verification and concurrent programming. We present a modular, flexible semantics for a significant subset of (close to) source code level Rust, based on the recent locally abstract, globally concrete semantics framework, separating local evaluation of expressions from their composition into concrete traces. The semantics is extended to model Rust's asynchronous programming features and Rust's most popular async runtime, Tokio. Based on our more abstract formalization, we establish the fairness of Tokio's scheduler. Further, we show the applicability of our semantics to deductive verification of Rust by providing soundness proofs for a Rust program logic.

Diffusion models provide expressive priors over trajectories, but adapting these priors to test-time constraints requires maintaining feasibility and consistency across successive control decisions. We introduce Persistent Particle Planning (P3), a sequential Monte Carlo framework for diffusion control that maintains a weighted population of candidate plans across replanning steps. At each control step, P3 shifts and partially re-noises the candidate trajectories, refines them under the latest observation, and uses constraint-aware weighting and resampling to select among alternative continuations without retraining the diffusion model. We consider denoising and replanning as one Feynman–Kac particle system and analyze it under an idealized repair. We prove that the re-noising depth controls how reliably a kept plan stays on its route, and that keeping a rare, well-separated route takes far fewer plans than rediscovering it by sampling from scratch. Experiments under multiple test-time constraint configurations show that population reuse reduces route switching and improves success without constraint violations. Because P3 refines earlier plans instead of redrawing them, it also needs fewer denoising iterations per replan. On maze-navigation tasks, it plans faster than both regenerated populations and methods that correct a single sampled plan by constrained optimization. Code and pretrained models are available at https://github.com/p3-username/p3-anon.

AI agents built on large language models (LLMs) run shell commands, read and write files, and reach the network, typically with their user's privileges. However, what an agent does during an execution is difficult to understand: tests assert on the result, and the agent's trajectory records only what the agent reports about itself, which may omit behavior executed by its subprocesses. We present AgentSpy, an approach that observes an agent from outside the agent. AgentSpy runs the agent in an isolated environment, configured by a declarative specification, and records the system calls and network traffic of the agent and of every process it executes. Based on this monitoring, AgentSpy supports two families of analyses: conformance analyses, which measure obligations, i.e., what an agent execution should do, and safety analyses, which check prohibitions, i.e., what an agent execution must never do. We instantiate one analysis of each family. The reliability analysis uses rules to summarize each run by the environment resources the agent uses: the commands it executed, the files it accessed, and the hosts it contacted. The security analysis applies deterministic rules to the system calls of an execution. For reliability, we evaluated AgentSpy on 77 tasks with the codex harness and three recent LLMs, executing each task three times. Sets of repeated runs of the same task are more similar than sets that include runs of another task in 92.2% of the comparisons. Among tasks for which all three runs pass outcome-based tests, the agent performs task-unrelated activities in 18% of the cases, reads the grading files in 7%, and does not use the developers' guidance in 17%. For security, the generic rules of AgentSpy detect four of five attack categories we considered, with no false positives across 50 runs.

Metastability separates apparent equilibration from true convergence: a dynamical system can remain near an intermediate state long before reaching its stationary state. In quantum dynamics used for optimization, such a plateau can delay progress toward a target concentrated near the global minimizers of a nonconvex potential. In this paper, we initiate the study of the general mechanisms by which coherent evolution and dissipation shape metastability, using quantum Langevin dynamics (QLD) as a concrete example [CLW+25], where the objective function is encoded in the potential and coherent transfer between wells interacts with tunable dissipation. To quantify this interaction, we solve a two-state model precisely and derive a spectral-separation criterion in terms of tunneling frequency, dissipation strength, and damping imbalance. Extending the analysis to multi-state models, we show that sufficiently weak tunneling between wells can separate slow population transfer from fast relaxation within each well. In addition, we establish conditions under which changing the structure of dissipation shortens the time needed to approach a prescribed target state. Numerical experiments on double-well and multi-well potentials demonstrate these improvements while keeping the target and accuracy fixed and accounting for the time spent on control.

While AArch64 CPUs are becoming strong market contenders, their software ecosystem lags behind the mature x86-64 environment, hindering the adoption of the new architectures and impacting user experience. Binary translation bridges this divide by converting binary code from one architecture (e.g., x86-64) to run on another (e.g., AArch64), allowing legacy software to benefit from modern hardware's performance and energy efficiency advantages. Current translation methods are typically either dynamic, which adds significant runtime overhead, or static, which struggles with reliability due to the inherent complexities of binary analysis. This paper introduces a new static, assembly-to-assembly translation paradigm that transforms binary code ahead of execution, generating portable, efficient native-like binaries that run on AArch64 devices without runtime frameworks. Benefiting from recent breakthroughs in large language models (LLMs), we provide a practical and automated translation engine that produces high-quality code with minimal human intervention. To ensure correctness, we introduce a crucial verification step, where we split the assembly code into simplified snippets, enabling efficient and scalable semantic verification. Our evaluation shows that this approach significantly outperforms existing open-source solutions with a large margin, producing binaries with near-native performance. Furthermore, it shows substantial improvements over the leading industrial translator, ExaGear, illuminating a promising new direction for cross-architecture binary translation research.

Molecular dynamics simulations proceed by integrating the Langevin equations over many small femtosecond timesteps. This poses a challenge for estimating ensemble properties and transition dynamics that occur on much longer timescales. We introduce Langevin Flow Maps, which extend machine-learned force-fields to additionally learn the stochastic Langevin integrator. We show that Langevin Flow Maps enable large-timestep molecular dynamics and recover accurate dynamical properties of the system, while running an order of magnitude faster than current machine-learned force fields. Further, by training on a diverse molecular dataset, we demonstrate a path towards transferable Langevin Flow Maps.

Early development unfolds in caregiver-infant dyads, where infants' sensorimotor streams are shaped by physical contact and face-to-face interaction. Yet developmental robotics simulators commonly model infants in isolation, limiting the study of caregiver-mediated experience. We present a caregiver-enabled extension of the Multi-Modal Infant Model (MIMo) in MuJoCo that turns MIMo into a controllable platform for replaying dyadic interaction and generating dense infant-perspective observations. The system provides (i) an articulated caregiver model compatible with MIMo morphologies, parameterized from anthropometrics and optionally resized to a recorded caregiver; (ii) a workflow to replay naturalistic caregiver-infant holding and soothing interactions; and (iii) logging and visualizing the infant's first-person multimodal experience (we show touch and vision). We showcase the tool on touch by introducing origin-aware contact logging that disambiguates self-, caregiver-, and environment-generated contact and supports aggregation into touch-rate statistics comparable to manual coding. While we showcase tactile analysis, the platform is intended more broadly as a generator of multimodal dyadic datasets (touch and egocentric vision) for modeling the development of social interaction.

Latent world models based on Joint-Embedding Predictive Architecture (JEPA) are deterministic by design. While successful in fully observable scenarios, this paradigm breaks down when past observations and actions lead to multiple plausible future possibilities, e.g., due to occlusion. We introduce EpicWorldModel, a framework to train stochastic JEPAs for environments and tasks with inherent uncertainty under partially observability. We jointly train the EpicWorldModel predictor with its latent representation space to directly predict multiple potential future states using a flow-matching objective, when the goal-relevant scene content is absent from the conditioning history. We show that flow predictive variance, motivated by its relation to an upper bound on predictive entropy, serves as a useful exploration guidance for planning. By incorporating this uncertainty signal into Cross-Entropy Method (CEM)-based planning, our approach balances goal-reaching with exploration of uncertain regions where occluded goals are most likely to be located. We demonstrate the effectiveness of EpicWorldModel through a series of latent planning experiments with the best or on-par performance across tasks, showing up to 22% empirical improvement in success rate over LeWorldModel.

We study the time evolution of the Operator Space Entanglement Entropy (OSEE) of one-dimensional number-conserving free fermions subject to a random space- and time-dependent evolution. It is known that, in free-fermion systems without randomness, the OSEE of local operators grows at most logarithmically in time. Here we show that, in the presence of randomness, the OSEE of local spin operators involving a Jordan-Wigner string grows as $\mathcal O(\sqrt t)$, for all values of the Rényi index. We establish this in two models: a Brownian Hamiltonian and a brick-wall matchgate circuit. We also consider the OSEE of the evolution operator, and we show that it also grows as $\mathcal O(\sqrt{t})$, in contrast with the linear growth generically expected without randomness. We analytically compute the average entropies of the two operators by mapping the quantum models to the symmetric simple exclusion process, relying on an assumption of self-averaging which we support numerically. We also consider two other types of randomness: random evolution in time that keeps spatial translation invariance, and quenched disorder. In the former case, we find that the OSEE also grows as $\sqrt{t}$. In the latter case, we find that the entropy grows as $\mathcal O( \log \log t)$ for Hamiltonian evolution both for local operators with a Jordan-Wigner string and for the evolution operator, whereas it remains bounded at all times in the brick-wall matchgate circuit.

Vision-language-action (VLA) models currently demonstrate strong performance in a wide range of real-world robotics tasks. However, they often still lack the generalization ability to handle large visual out-of-distribution shifts. Post-training of VLAs with reinforcement learning (RL) has been shown to benefit robustness, but significant room for improvement remains. In this work, we systematically study the effect of image augmentation on VLA post-training. We find that it is crucial to augment only the critic module during RL updates, while leaving the actor's input clean during both rollouts and updates. For $\pi_{0.5}$ and GR00T N1.5 this raises out-of-distribution success on LIBERO-Plus by $7.8$ and $10.0$ points respectively, while augmenting the actor collapses training entirely. We investigate a range of augmentation types and strengths, and provide practical recommendations for improving generalization in VLA post-training.

Composed image retrieval (CIR) aims to retrieve a desired target image from a query consisting of a reference image and a modification text. This task exhibits an unusual representational asymmetry: the modification text specifies a transition from the reference state, whereas retrieval candidates depict completed target states. This creates a representation mismatch for zero-shot methods that query pretrained vision-language spaces directly with transformation-oriented language. We study this mismatch and reformulate zero-shot composed image retrieval as target-state reconstruction followed by retrieval. We instantiate this formulation with ASAP-CIR, a training-free framework that reconstructs a static target representation using a frozen multimodal large language model (MLLM). The representation combines multiple holistic descriptions with a variable set of importance-weighted atomic semantics, thereby preserving both overall target identity and fine-grained visual constraints. Retrieval then integrates holistic state alignment, atomic constraint grounding, and calibrated target-state evidence aggregation. A controlled text-only diagnostic shows that target-side static query formulations achieve more reliable retrieval than dynamic composed query formulations, particularly when source-state semantics must be suppressed or transformed. Experiments on FashionIQ, CIRR, and CIRCO further characterize the effectiveness and limitations of this representation principle, with the clearest gains on the multi-target CIRCO benchmark. These results show that how composed intent is represented before retrieval is a consequential design choice, distinct from the choice of retrieval backbone itself.

We study transmission of a Greenberger-Horne-Zeilinger (GHZ) state over nonunitary quantum multiple-input multiple-output (QuMIMO) channels using finite-energy Gottesman-Kitaev-Preskill (GKP) encoding. A coupled-mode model describes coherent mixing, mode-dependent attenuation, and Gaussian noise. The transceiver combines programmable passive meshes, quantum-limited MMSE gain, local GKP recovery, and joint classical syndrome decoding. We optimize complex-Givens lattices using instantaneous channel-state information (CSI), current singular values with channel statistics, or statistics alone. Simulations show that joint lattice shaping outperforms product GKP processing. Under coherent-basis drift, full instantaneous CSI improves GHZ-state preservation, whereas singular-value conditioned and fully statistical designs perform similarly because neither resolves the current singular frames. These findings identify singular-frame geometry as a critical resource beyond attenuation knowledge where joint GKP processing adapts the logical lattice and receiver to anisotropic bosonic noise rather than merely inverting multimode mixing.

Trust is crucial in the digital world, particularly in the age of systems based on artificial intelligence. Traditionally, system correctness is demonstrated by proving that the system satisfies the required properties. However, digital systems involve humans, who can make mistakes, break rules, or even intentionally cheat the system. Technical devices may also malfunction or cause errors. Therefore, external stakeholders need to trust a system throughout its operation. To address this challenge, we propose the concept of continuous trust. In this paper, we present a detailed explanation of this concept, based on a running example of a business's refund department. The paper concludes with a discussion of related work, such as compliance checking and certification, as well as an outlook on further questions regarding the formalization and the necessary tool support.

We show that whenever $G$ is a finitely generated group with the rapid decay property (property RD), then the norm on the reduced group C*-algebra $C^*_r(G)$ is computable if (and only if) $G$ has solvable word problem. We also give an example of a finitely presented group $G$ with solvable word problem for which the norm on $C^*_r(G)$ is not computable.

Large-language-model release pipelines increasingly combine commitments, signatures, provenance records, and heterogeneous verification backends. Yet validating every submitted object does not establish that a release realizes every requirement of its registered transformation. An untrusted realization proposer may omit a required relation, propose an unauthorized evidence-sharing assignment, or bind valid evidence to the wrong object. This verification-boundary failure is termed Semantic Omission under Valid Evidence (SOVE). Hybrid Cryptographic Relation-based Semantic Plan Verification (H-CRSPV) is a replicated admission layer that enforces required-set completeness before release consumption. Before evidence selection, an authorized registration entity commits an authenticated authority record. Validators independently derive the required-obligation multiset, check exact entry-occurrence coverage, validate proposal-induced evidence sharing, and bind admissible groups one-to-one to keeper-resolved objects. Evidence remains provisional until the challenge window closes and an atomic finalizer activates the release. The analysis establishes structural exactness and conditional semantic guarantees under explicit assumptions. Across 36 omission artifacts, submitted-object validation accepts all 36 because every submitted object passes its backend-specific verifier. H-CRSPV rejects all 36 while accepting all six honest releases. The prototype also validates six restricted source-to-RelationIR bindings and rejects all 60 tested mutations. In a continuous four-validator Qwen2.5-1.5B workflow, the same authority record remains fixed across three legal releases with different post-registration availability states. These results show that per-object validity does not establish complete, correctly bound, and finalized release evidence.

Let $\mathfrak g$ be a simple Lie algebra of classical type. We construct a $\mathbb Z[1/2]$-integral form of the Yangian $Y(\mathfrak g)$ generated by divided powers of the positive and negative Drinfeld generators, and establish its triangular decomposition and PBW basis. Under the loop filtration, its associated graded algebra is the current algebra part of Garland's integral form. As applications, over an algebraically closed field of characteristic $p>3$, we identify the small Yangian with the restricted Yangian and classify their finite dimensional simple modules in terms of restricted Drinfeld polynomials.

Computational pathology has the potential to improve clinical outcomes through a demonstrated increase in diagnostic and prognostic accuracy. However, the development and validation of deep learning algorithms still require annotated data, a costly procedure involving expert pathologists who already face critical workforce shortages. Existing coreset selection methods to optimize annotation efforts currently all rely on hyperparameters tuned on natural-image benchmarks that do not transfer to histopathology and are cumbersome to use in clinical practice. In this study, we present GCcore, a novel label-free coreset selection method that embeds every image of a dataset with any pathology foundation model and greedily selects the samples that collectively maximize the global coverage of the embedding space. The proposed method provides a lower-bound guarantee on the global coverage of the returned coreset for any coreset size, while being completely hyperparameter-free and deterministic. We demonstrate GCcore's superior performance over 14 baselines including state-of-the-art methods across 10 tasks and datasets spanning whole slide image classification, tile classification, and tissue segmentation, where it ranks first on six and within the top three on nine, while also demonstrating how existing methods can shift by up to five rank positions depending on their hyperparameter settings. Code is publicly available at https://github.com/OncoAI-ULBHUB/GCcore.

Answering a question of Thomas Sinclair, we show that the standard presentation of the reduced group C*-algebra $C^*_r(G)$ of an infinite hyperbolic group $G$ is computable. The main ingredient in proving this result is showing that the Gromov boundary $\partial G$ of $G$ admits a computably compact metric presentation for which the natural action of $G$ on $\partial G$ is computable.

We study the high-frequency Fourier coefficients $C_n$ of critical Gaussian multiplicative chaos $M$ on the circle, for logarithmic covariances with an arbitrary admissible smooth, possibly nonstationary remainder. For every fixed finite set of integer offsets, the corresponding vector of neighboring coefficients, multiplied by $\sqrt{\log n}$, converges stably relative to the full Gaussian field without additional centering. Conditionally on that field, the limit consists of Fourier coefficients of one symmetric complex stable random measure of index one with control proportional to $M$. For every $s>1$, the normalized modulated measures also converge stably in $H^{-s}(\mathbb{T})$ to this conditional stable noise. Unconditionally, the same joint approximation remains valid for offsets of size at most $(\log\log n)^{1/64}$ for every fixed admissible covariance. The scalar limit is an isotropic Cauchy mixture; the joint limiting law retains the spatial distribution of $M$, not only its total mass. We also prove an almost-sure integral test along the entire integer frequency sequence. For a class of regular deterministic gauges $H$, the limsup of $\sqrt{\log|n|}\,|C_n|/H(\log\log|n|)$ is zero or infinity according as $\int^\infty dx/H(x)$ converges or diverges. The proof combines index-one compensation with relative excursion estimates and transfers these conclusions from the canonical model by smooth covariance comparison.

The nuclear emulsion method is applied to study binary dissociation leading to the formation of $\alpha$-particles from $^{16}$O nuclei at 3.65 GeV per nucleon and $^{22}$Ne nuclei at 3.22 GeV per nucleon. The invariant mass distribution for the $^{16}$O$\to$$^{12}$C(0$^+_1$)+$\alpha$ channel indicates decays of the $^{16}$O(0$^+_6$) excitation, whose probability exceeds the estimate for the $^{12}$C(0$^+_2$)+$\alpha$ channel by a factor of 20. The invariant mass distribution for $^{22}$Ne$\to$$^{18}$O+$\alpha$ exhibits a peak corresponding to $^{22}$Ne excitations around 11.6 MeV.

We demonstrate sub-Doppler cooling of rubidium ($^{87}$Rb) atoms using a compact, single-beam configuration based on a conical hollow mirror integrated with anti-Helmholtz coils. The conical geometry converts a single incident beam into an effective six-beam magneto-optical trap (MOT) configuration while focusing the light towards the cone axis, thereby reducing the required laser power. The number of atoms in the conical MOT is maximized at an incident intensity of 6.5 $\mathrm{mW/cm^2}$ in our conical MOT, several times lower than in conventional six-beam MOTs. Under optimal conditions, we trap $\sim 1 \times 10^7$ $^{87}$Rb atoms. We obtain sub-Doppler temperature of $T_r=13.4\pm 1.2\ \mathrm{\mu K}$ and $T_z=106\pm 11\ \mathrm{\mu K}$ by polarization gradient cooling (PGC), revealing a strong anisotropy that we attribute to the polarization pattern created by the conical mirror. In addition, a modified release-and-recapture measurement in the MOT yields a radial temperature $T_r = 1.20\pm 0.02$ mK, approximately five times higher than that expected for comparable light-shift parameters in a six-beam MOT, in agreement with previous simulations. Our results establish conical hollow mirrors as a practical and power-efficient platform for compact cold-atom sources and integrated quantum sensors.

With the rapid development of artificial intelligence, the emergence of various Large Language Models (LLMs) has created a rich model ecosystem. However, this also brings a key challenge: how to select the optimal model for a specific user query. LLM routing addresses this need by dynamically assigning queries to the most suitable expert in the pool of candidate models. However, existing routing frameworks often simplify this process to a standard classification task; thus, a critical vulnerability is exposed when multiple candidate models correctly answer the same query. We formalize this capability overlap as routing noise, which misleads the router with arbitrarily correct candidate models, ultimately leading to routing collapse (a severe decline in generalization ability on unseen tasks). To address this problem, we propose a novel Cluster-Aware Soft-Labeling Routing (CASLR) framework. CASLR shifts the evaluation paradigm from the success of a single query to macro-domain consensus by replacing traditional one-hot vectors with a masked softmax mechanism. Specifically, for experts who answer incorrectly, we penalize their target probability to zero; for the remaining candidates, we directly compute continuous fine-grained soft labels based on their global clustering utility scores. We then use these refined soft labels to supervise a lightweight router. Specifically, the framework not only demonstrates superior accuracy on multiple benchmarks, but also outperforms Llama-3.3-70B-Instruct by 7.80% in overall average performance. Furthermore, the extremely low routing inference latency of only 1.13s further confirms that CASLR can achieve efficient system scheduling with almost zero additional overhead, while ensuring high response quality.

A prior longitudinal study of silent failures in a production LLM agent runtime (arXiv:2606.14589) found that about 70% were discovered by a human looking at the product as a user while thousands of tests and governance checks stayed green, and posed mechanizing part of what the human eye does as an open problem. This paper reports our attempt. We built an automated user-viewpoint observer targeting the most dangerous class, fail-plausible failure, in which an internal error becomes fluent, plausible output to the user. It is a two-layer pipeline: five deterministic signals distilled from incident postmortems escalate to an LLM judge whose verdicts must cite verbatim evidence or be discarded. Ground truth is 24 labeled production postmortems with explicit honesty boundaries: 16 of 24 are structurally invisible to any content-reading observer, and we say so. Offline, the deterministic layer achieves 6/6 regression detection with 0/4 false positives, each detector proven load-bearing by sabotage; held-out recall on novel patterns is 0/4. It is, so far, a regression engine. Deployment followed pre-registration: shadow mode caught and retired one systematic false positive on its first run, the 26-day shadow window then ran clean, flip criteria were fixed before the shadow data was read, and the analysis protocol was frozen before the enforcing-mode window opened. That window (12 observed days) fired zero verdicts; per the pre-registered path we report live precision as undefined rather than narrating quiet as success. The observer also produced ten silent failures of its own, confirming that the judge inherits the taxonomy it judges. We release the corpus, detector, and scorecard as a runnable bench: mechanization today retires the human's regression scanning so the eye can specialize in novelty; prediction remains open but is now measurable under frozen rules.

We present the first search for the diffuse supernova neutrino background utilizing coherent elastic neutrino-nucleus scattering with the PandaX-4T liquid xenon detector at the China Jinping Underground Laboratory. Based on a total exposure of $1.54~\mathrm{tonne \cdot year}$ from the commissioning and first science runs, we observe no significant event excess over the expected background. We therefore set an upper limit of $(2.27 - 2.43) \times 10^3~\mathrm{cm^{-2}s^{-1}}$ per flavor at 90% confidence level on the $\nu_x$ flux (each of $\nu_\mu, \bar{\nu}_\mu, \nu_\tau, \bar{\nu}_\tau$) of the diffuse supernova neutrino background for incident neutrino energies above 19.3~MeV. The result is comparable to the current world best limit.