Abstract The proposed enhancement method (SGLC) effectively addresses the challenges of blurred details, low contrast, and noise interference in infrared weak-texture images. The approach involves a multistep process. First, guided filtering is employed to decompose the original image into a base layer sub-image containing low frequency information and a detail layer sub-image with high frequency information. The base layer image is then processed using the LoG-CLAHE algorithm, which enhances contrast and brightness while suppressing noise and sharpening edges, thereby improving texture information. The detail layer image undergoes an improved gamma correction to prevent detail loss in low contrast regions. Finally, the base layer and detail layer images are fused using weighted factors to generate the enhanced infrared weak-texture image. Experimental results across three different scenarios demonstrate the effectiveness of the proposed algorithm. Significant improvements are observed in mean local information entropy, average gradient and fuzzy comprehensive evaluation compared to other methods. The enhanced infrared weak-texture images exhibit improved clarity and contrast, with prominent texture details, validating the efficacy of the SGLC approach for infrared weak-texture image enhancement.

Abstract Background: Although the flipped classroom is increasingly used in nursing education, previous research has focused mainly on outcomes such as academic achievement and satisfaction. Less is known about how nursing students experience flipped learning as a learning process, particularly in relation to learner agency, interaction, and the coexistence of growth and burden. Aim: To explore the meaning of undergraduate nursing students’ learning experiences in a flipped classroom implemented in a community health nursing course from a phenomenological perspective. Design: A qualitative study using descriptive phenomenology. Settings: A community health nursing course in an undergraduate nursing programme. Participants: Forty-eight undergraduate nursing students who participated in the flipped classroom course and voluntarily submitted reflective journals. Methods: Data were collected from individual reflective journals written between January 1 and January 30, 2024. The data were analysed using Colaizzi’s phenomenological method. Results: Four themes were identified: learning that begins even before entering the classroom, becoming aware of oneself as an active learner, relational experiences in which learning expands through speaking and questioning, and ambivalent experiences in which temporal and emotional burdens coexist. Students experienced the flipped classroom as a continuous and meaningful learning process shaped by preparation, participation, interaction, and responsibility. At the same time, they perceived that the learning process involved both educational growth and psychological and temporal burden. Conclusions: The flipped classroom was experienced not simply as an instructional technique, but as a pedagogical design that reshaped when learning began, how learners participated, and how meaning was constructed through interaction. These findings suggest that flipped learning in nursing education should be designed to support active and meaningful engagement while carefully managing workload, emotional burden, and fairness in collaborative learning.

Within general relativity, black holes are widely regarded as the ultimate benchmark for compactness in the Universe. Recently, however, neutron star models have been constructed in a higher-curvature theory -- quasi-topological gravity (QTG) -- whose compactness can exceed the black-hole limit~\cite{LD19666}. Here we present a detailed analysis of both the equilibrium structure and radial stability of such configurations in QTG. By examining several representative equations of state and different values of the gravitational coupling constant, we find that in the high-central-density regime the compactness exceeding the black-hole bound exhibits a universal behavior in QTG. We further show that QTG corrections grow increasingly significant at large central densities and can stabilize configurations that are radially unstable in general relativity over a broad parameter range. These results establish ultra-compact neutron stars in QTG as theoretically viable strong-field configurations and provide a foundation for further investigations of their dynamical and phenomenological implications.

Cascade attacks in LLM multi-agent systems (MAS) arise when adversarial influence propagates across agents and leads to escalated system-level failures through complex agent interactions. Detecting such cascades is challenging, as their signals are distributed, tightly coupled across interaction channels, and often appear plausibly benign locally but may unfold quickly either within a single turn or gradually across multiple turns. Existing defenses, being largely local and text-centric, fail to capture such cross-channel, temporally coordinated dynamics of cascade propagation. Therefore, we propose CASPIAN, the first framework that provides a unified, cross-channel causal analysis of cascade behavior in LLM-MAS through online monitoring of dynamic influence propagation across agents. CASPIAN models multi-agent interactions using a unified, dynamic causal influence matrix across channels, estimated efficiently via a late-interaction conditional transfer entropy (LI-CTE) formulation, thereby enabling the detection of cascade onset from emergent system-level structure rather than isolated anomalies. It further performs online causal attribution, identifying the origin, bridge, and amplifier agents driving the cascade and reconstructing its principal propagation pathways, capabilities not supported by existing methods. Across diverse multi-agent frameworks and benchmarks, CASPIAN consistently outperforms semantic guardrails, LLM-based judges, and graph-based anomaly detectors in both detection accuracy and early cascade identification while operating with sub-1% relative overhead latency. These results demonstrate that unified cross-channel causal modeling is essential for reliably detecting and understanding cascade failures in LLM multi-agent systems.

World models are widely explored in embodied intelligence, yet they typically predict distinct evolutions of the world and the ego within a single stream, where the world captures persistent instruction-agnostic scene regularities and the ego captures robot-centric instruction-conditioned dynamics. This world-ego entanglement leads to a degradation in long-horizon embodied scenarios, particularly in hybrid tasks with interleaved navigation and manipulation behaviors. In this paper, we introduce \emph{World-Ego Modeling}, a new conceptual paradigm that decomposes future evolution into world and ego components. We define the world-ego boundary from three perspectives, i.e., motion-, semantic-, and intention-based views, and analyze three disentanglement strategies with post-, pre-, and full disentanglement. Further, we instantiate this paradigm as the World-Ego Model (WEM), a unified embodied world model that couples an implicit separate world-ego planner with a cascade-parallel mixture-of-experts (CP-MoE) diffusion generator. To enable rigorous evaluation, we further construct HTEWorld, the first benchmark for long-horizon world modeling with hybrid navigation-manipulation tasks, providing 125K video clips (over 4.5M frames) with fine-grained action annotations and 300 multi-turn evaluation trajectories (over 2K instructions). Extensive experiments show that WEM achieves state-of-the-art performance on HTEWorld while remaining competitive on existing manipulation-only benchmarks.

We present a new accurate analysis of the $^3$He$(p,e^+ν_e)$${}^4$He (''hep'') reaction at astrophysical energies. The S-factor is computed using a state-of-the-art method to calculate the four-nucleon scattering and bound-state wave functions (the hyperspherical harmonic expansion), and by using nuclear interactions and accompanying electroweak nuclear currents obtained within the chiral effective field theory framework. Our analysis includes a detailed examination of the theoretical uncertainties coming from two different sources: the truncation of the interaction and current chiral expansions, and the model dependence. Our recommended final theoretical value for the hep S-factor at zero energyis $S(0)=(8.7\pm 0.9)\times 10^{-20}$ keV b. We provide also the energy spectrum of the outgoing hep positrons which may be measured in future experiments. We include also an analysis of the ''sister'' reaction $^3$He$(n,γ)$${}^4$He (''hen'') at low energies, showing that the calculation well reproduce the total cross section from thermal energies to few MeV, validating our results on the hep reaction.

Mathematical models are invaluable for understanding and predicting how biological systems behave, although their construction requires specifying mechanisms and relationships that are often not perfectly known. In the presence of multiple competing models, model uncertainty should be accounted for when performing inference based on available data. Bayesian model selection is a framework for testing mechanistic hypotheses and generating predictions under model uncertainty, which generally requires computation of the model evidence. In this work, we investigate the reliability of evidence computation methods when parameter non-identifiability -- the inability to distinguish between parameter values given available data -- is present, and find that deterministic evidence approximations can produce misleading model selection results because their underlying assumptions are violated. We propose a novel implementation of adaptive multiple importance sampling for evidence estimation, and demonstrate its robustness against non-identifiability. We use ecological case studies to demonstrate how simple model selection methods fail to produce accurate results, whereas our method yields model selection results that are comparable to those obtained by Markov chain Monte Carlo methods at substantially lower computational cost. Given the pervasiveness of parameter non-identifiability in mathematical biology, this work provides a practical approach to reliable model selection in the presence of poorly identified parameters.

In this paper, we establish Berry-Esseen-type bounds for federated linear stochastic approximation (LSA). Our results provide the first federated Gaussian approximations for LSA that explicitly capture communication-computation trade-offs and heterogeneity-aware error terms, quantifying the effects of local step size, number of local updates, and heterogeneity on convergence rates. We present results for both (i) constant step size regime and (ii) decreasing step size with an increasing number of local iterations, recovering the recent rates of Bonnerjee et al. [2025] as a special case. As a primary application of our results, we develop an online multiplier bootstrap procedure for inference on the last iterate, which avoids explicit estimation of the asymptotic covariance matrix, and obtain non-asymptotic validity guarantees for this procedure.

The generation of immersive and navigable 3D environments is increasingly prevalent with the growing adoption of virtual reality and 3D content. However, recent methods face a fundamental limitation: they cannot produce 3D worlds that simultaneously (i) are navigable over long-range spatial extents and (ii) cover the complete omnidirectional field of view ($360^\circ$ horizontally and $180^\circ$ vertically). To address this challenge, we introduce SphericalDreamer, a method for generating fully immersive and long-range 3D outdoor environments from textual prompts. Our approach is built on the generation of multiple panoramic images, which are subsequently lifted into 3D and fused together while maintaining visual and geometric consistency. SphericalDreamer produces highly detailed, fully immersive 3D environments, while substantially improving scale and navigability compared to prior approaches.

Robot guide dogs offer navigation assistance that greatly expands the independent mobility of the visually impaired, but their effective use requires subtle human-robot coordination that is difficult for users to learn from generic verbal instructions. To tackle this challenge, we present CANINE, an automated coaching system that trains users for interactive navigation with a robot guide dog, through personalized, adaptive verbal feedback. CANINE decomposes a complex coordination task into sub-skills and operates at two levels. At the high level, it decides what to train by tracking the learner's proficiency across sub-skills using knowledge tracing and prioritizing training on the weakest areas. At the low level, CANINE decides how to train each sub-skill by observing each human practice episode, using foundation models to infer the underlying causes of errors, and generating targeted verbal corrections adaptively. A controlled study with blindfolded participants, treated as a proxy population for quantitative evaluation, demonstrates that CANINE significantly improves both learning efficiency and final navigation performance compared to generic verbal instructions. We further validate CANINE through a retention study and an exploratory case study. The retention study shows lasting skill improvement after two weeks. The case study confirms CANINE's effectiveness in training a visually impaired user, while revealing additional design considerations for real-world deployment. Both are well aligned with the findings of the controlled study. Project page: https://cunjunyu.github.io/project/canine/