Papers

Cell Painting is a high-content morphological profiling assay widely used for phenotype-based biological inference, with mechanism of action (MOA) prediction as a central application. Existing approaches largely formulate Cell Painting-based inference as representation matching, assigning predictions from nearby reference perturbations in morphological feature space. However, retrieved neighbors are often noisy and partially misleading evidence due to batch effects, non-specific cytotoxicity, phenotypic convergence, and source-dependent variability. We reformulate Cell Painting-based MOA prediction as a calibrated evidence reasoning problem, where retrieved neighbors are treated as uncertain observations that must be evaluated, compared, and sometimes rejected before supporting a mechanistic conclusion. We propose PhenoAIR, a reliability-aware multi-agent framework that maintains a candidate-centric evidence memory and performs controller-guided refinement over phenotype- and mechanism-side evidence. PhenoAIR uses offline reference-set calibration to weight evidence by source reliability, phenotype stability, and mechanism-level confusion. We evaluate PhenoAIR on a benchmark constructed from JUMP Cell Painting profiles and annotations, covering controlled, realistic, and discovery-oriented open-world MOA prediction settings. PhenoAIR outperforms representation-matching and LLM-based baselines across all settings.

We study a finite-horizon reversible investment problem in which a risk-neutral firm adjusts capacity at a proportional purchase cost and a lower salvage value under multi-factor geometric Brownian motion. Via the singular control–optimal switching correspondence, the marginal value of capacity solves a family of parabolic double-obstacle problems. We prove existence, uniqueness and local Sobolev regularity of the strong solution, characterize investment, waiting and disinvestment regions by continuous, strictly separated free boundaries, and verify optimality of the reflected capacity process. Numerically, joint demand improvements shift both boundaries super-additively, 1.5–2.7 times as strongly at the disinvestment boundary, depending on factor correlation.

Breakthrough curves (BTCs) record arrival times, but anomalous transport models are often fitted with time-normalized spatial concentration profiles. Here, the effect of this mismatch on groundwater-transport upscaling is tested. Analytical limits and high-$q$ scaling were verified first. Six first-passage models were then fitted to particle arrivals from 36 two-dimensional digital aquifer runs. The set included exact ADE, $q$ and time-memory extensions, a joint memory-$q$ model, Gamma-CTRW, and a two-path inverse-Gaussian mixture. After the operator was corrected, memory-qFPT was not selected in any run. AICc selected the two-path model in 20 runs, Gamma-CTRW in 15, and q-FPT in one. Median downstream holdout log-RMSE was 0.904, 0.914, and 0.959 for the two-path, Gamma-CTRW, and memory-qFPT models, respectively. The memory-$q$ parameters also failed an independent plume-scaling test ($R^2=0.189$; slope $=-0.332$). Meanwhile, multimodal downstream BTCs increased from 0% in homogeneous media to 100% in the strongest heterogeneity class. Gamma-CTRW was also favored in an independent synthetic field. Observation consistency is therefore necessary but not sufficient for physical validation. In these simulations, pathway and waiting-time structure was supported more strongly than a unique memory-$q$ law.

In various special cases, we provide an alternative argument for equivalence between Kostant positivity of some involutions in type $A$ and the conjunction of the categorical Kåhrström's condition for these involutions with indecomposability of the images of the corresponding simple modules in category $\mathcal{O}$ under projective functors.

Let $G_m$ be a bounded Vilenkin group and put $R:=\sup_j m_j$. We prove the sharp uniform estimate $$\|K_n\|_1\leq \frac{17}{15\ln 2}\ln R,\qquad n\geq1,$$ for the Fejér kernels associated with the Vilenkin system. The constant $17/(15\ln2)$ is best possible. In the Walsh–Paley case $R=2$, the result reduces exactly to Toledo's sharp estimate $\sup_n\|K_n\|_1=17/15$.

Characterizing the properties of neuronal cell types is central to understanding brain function. Yet, biophysically realistic neuron models remain difficult to scale to large populations while preserving the richness of cellular dynamics and within-cell-type variability. We present a cell-type-specific neural-operator framework for scalable neuronal population modeling. Rather than training separate surrogate models for individual neurons, the proposed framework learns a shared operator across multiple modeled neurons from the same cell type, conditioned on biologically interpretable electrophysiological features. Relative to the original single-neuron NOBLE setting, learning across heterogeneous neurons requires a richer electrophysiological conditioning representation and a training objective designed to preserve diverse spiking dynamics. This enables efficient generation of large ensembles of somatic voltage responses while preserving variability across cells and model configurations. We evaluate the approach using membrane dynamics metrics, action potential waveform metrics, spike timing accuracy, input-output firing relationships, and electrophysiological feature distributions. Our framework reproduces subthreshold membrane dynamics, spike waveforms, and firing-rate responses across major cortical inhibitory neuron types, while accurately preserving spike timing despite the sensitivity of threshold-crossing dynamics. The predicted electrophysiological feature distributions show strong agreement with detailed simulator outputs and human experimental recordings, including high explained variability in properties such as inter-spike intervals, firing frequency, action-potential shape, and post-spike recovery.

Oscillatory phase coding (OPC) is widely observed in neural systems and has been implicated in cognitive functions such as memory and sequence processing. However, its computational role and neural substrates remain largely unclear. Here, we studied OPC by training recurrent neural networks (RNNs) with adaptation to perform sequence working memory tasks with variable delays. After training, we found that OPC emerges in the RNN spontaneously: the memorized items recur at distinct phases of a shared rhythm, with oscillation amplitude encoding item identity and phase encoding serial order. Ablation studies identified three key elements required for OPC: adaptation in neural dynamics, variable delay in memory retention, and flexibility in reading out information. Neuron-type analysis revealed the emergence of content and time cells in the trained RNN, and they process item identity and sequence order separately. Recurrent connectivity analysis revealed phase-organized connections among time cells and structured connections within digit-selective content-cell groups. Inspired by these findings, we proposed a neural circuit model which reproduces OPC through interactions between content units and an adaptive time ring. Together, this study reveals how OPC arises from network structures and task demands, offers insights into the implementation and computational mechanisms of OPC in biological systems, and inspires us to develop computational models for human-like cognitive functions.

Heat-related mortality varies among individuals under similar environmental conditions, yet environmental exposure and genetic susceptibility are typically modelled separately. We propose a signal-processing cascade linking satellite-derived land-surface temperature to sequence-dependent heat-shock response and mortality risk. The framework combines distributed-lag filtering of thermal exposure, nonlinear HSF1 activation, cross-correlation with an HSF1 position-weight-matrix kernel, and Hill-model mapping of sequence-dependent binding changes to regulatory response. Applied to 52 heat-response genes and 157 naturally occurring variants, the framework identifies 21 heat-shock-element arrays across 17 genes. Variants in canonical arrays are significantly enriched for predicted binding disruption ($p=3\times10^{-4}$). Propagating these effects through the cascade yields a peak predicted mortality-risk multiplier of $1.68\times$ at $40$–$42\,^{\circ}\mathrm{C}$. The framework demonstrates how environmental and genomic signals can be integrated to model inter-individual variation in heat-related risk.

Many researchers believe that human lifetime on the interval (30,80) years is described by the Gompertz-Makeham distribution. In this work, the task is set to test this hypothesis. We consider a number of distributions: Gompertz-Makeham, Gompertz-Weibull, Logistic Kannisto-Makeham, Logistic Beard-Makeham and Logistic Beard-Weibull, some of which are partially known, for which comparative calculations on real mortality data are performed using the $\chi^2$ criterion. The data were selected for three countries: England & Wales, Norway and Sweden, cohorts born in 1900. As it turned out, on the interval (30,80) years the Gompertz-Makeham distribution is far from being the best among those listed above. In addition, for the same data, a similar problem was studied for the interval (80,120) years.

In this paper, we introduce and study the notion of character cyclic amenability for Banach algebras, which combines the cyclicity condition on derivations with the character-dependent module structure arising in character amenability. We establish several basic properties of this notion and investigate its behavior under standard Banach algebra constructions. In particular, we study character cyclic amenability for opposite algebras, projective tensor products, l1-direct sums, quotient algebras, dense homomorphic images, and theta-Lau products. We also provide examples illustrating the relationship between character cyclic amenability and other amenability-type properties, including examples showing that character cyclic amenability does not imply cyclic amenability. Finally, we characterize character cyclic amenability for module extension Banach algebras in terms of phi-cyclic derivations and derivations associated with the underlying Banach module.

Referring expression segmentation (RES) aims to localize targets at the pixel level guided by human instructions. However, current models predominantly follow a text-centric paradigm that treats discrete text as the primary interaction medium for these human instructions. This assumption is especially limiting in unmanned aerial vehicle (UAV) applications, where rapid response, hands-free operation, and natural communication are often critical. In this paper, we reconstruct UAV referring expression segmentation through an audio-centric perspective. Grounded in information theory, we demonstrate that raw audio provides a higher upper bound for extracting rich semantic features than transcribed text. To achieve this, we construct an audio-driven UAV referring expression segmentation benchmark named Aerial-AudioRES by augmenting VisDrone with synthesized speech, forming "audio-text-image-mask" quadruples. Specifically, we employ an ambiguity filtering mechanism to ensure referential uniqueness, and incorporate eight distinct timbres alongside 4 realistic noises to simulate authentic environments. To drive the paradigm shift from text-centric to audio-centric, we propose the end-to-end Audio-centric Referring Expression Segmentation Network (ARES-Net). Distinct from discrete text-visual matching, our approach learns to establish a direct soft alignment between continuous acoustic manifolds and multi-scale visual representations. Comprehensive experiments show that our method obtains a peak accuracy of 64.23% in oIoU and demonstrates robust generalization on a collected real-human test set. Furthermore, it significantly reduces total interaction latency to one-third of the text-based method. Our code is available at https://github.com/LemonGauss/ARES-Net.

Object detection is a critical application of computer vision, yet models trained uniformly on bounding boxes lack the geometric precision to be fully effective in inference settings, especially at high IoU thresholds. One common approach is the introduction of polygon-supervised localization, where a model is trained explicitly for localization but with access to polygon masks during training. Among such methods, partial polygon supervision is viable due to the higher annotation costs of developing fully supervised training datasets. However, this introduces a further challenge: in object classes of relatively low shape complexity, the addition of training masks and subsequent segmentation loss may obscure box localization, often worsening model performance despite access to deeper geometric information. To address the practical limitations of both full and partial polygon supervision, we introduce a mathematical framework for quantifying per-class geometric complexity by combining information theory and statistical shape analysis. With this framework, we develop Entropy-Aware Geometric Learning Enhancer (EAGLE), a lightweight, post-training model head that learns geometric information for the localization model, improving accuracy via edge refinement and localization rescoring using entropy-informed updates. Empirical analysis across MS COCO and Pascal SBD demonstrates the entropy framework meaningfully predicts shape reconstruction error and accuracy differences among mask-derived and standard box predictions, while the implemented EAGLE framework achieves consistent improvement on already SOTA detection frameworks (+0.9 mAP50-95 points on MS COCO) at minimal observed inference overhead.

Reconstructing sub-millimeter detail of small objects underpins tasks such as circuit-board defect detection, specimen digitization, and archiving. 3D Gaussian Splatting (3DGS) makes photorealistic reconstruction possible from posed images alone, but macro optics violate two conditions it depends on. The field of view (FOV) and the depth of field (DoF) shrink to millimeters, so viewpoint error can take a sample out of frame, and only a thin focal plane is sharp in frame. Existing automated macro-capture rigs handle the shallow DoF with per-view focus stacking, but typically rely on classical photogrammetry and offline, human-in-the-loop processing. We present MicroSplat, which generates a 3DGS using a low-cost consumer microscope on a manipulator arm. After a single up-front object-localization step, the system runs unattended. It performs calibration, plans collision-free viewpoints, captures and fuses a focal stack at each viewpoint, segments the object across all views, and optimizes a 3DGS model warm-started with the arm's own poses. We further compare against blur-aware 3DGS methods, which infer sharp scenes from defocused images, and find that measuring sharpness with a controlled focal sweep outperforms inferring it. We provide real and synthetic datasets (including robot and camera parameters) and the fitted 3DGS reconstructions.

We present RhinoNet, a simplicial complex-based point cloud neural network, that creates a geometry-preserving embedding of high-dimensional point clouds. RhinoNet takes collections of datasets, also called data cohorts, and processes each dataset (point cloud) within the cohort as a simplicial complex. These simplicial complexes have datapoints as vertices and edges learned by Vietoris-Rips filtration. Using simplicial wavelets and MLP layers, RhinoNet computes a vector representation of each dataset in a latent space produced by an autoencoder. The latent space is optimized such that an optimal transport distance between pairs of point clouds is preserved between the latent representations. We apply RhinoNet to single-cell datasets collected on patients—a common source of such high-dimensional data. RhinoNet creates a patient level representation, known in biological contexts as a phenoscape. We establish that RhinoNet generates organized phenoscapes, whereas other methods such as static maps do not. Next we show that we can generate entire point clouds by sampling and decoding this phenoscape. We show two modes of generation via 1) latent flow matching and 2) geodesic interpolation in latent space, followed by a decoding back into the point cloud space. RhinoNet achieves state-of-the-art performance on single-cell and spatial data cohorts relative to tested baselines. Overall, RhinoNet lifts the concepts of representation learning and generative modeling to the point cloud or population level and defines a neural model for phenoscaping.

This review presents a concise and unified treatment of Wilson lines and Wilson loops in continuum gauge theory, emphasizing their interpretation as parallel transport and their status as gauge covariant transporters and gauge invariant loop observables. Starting from the problem of comparing charged fields at separated points, we derive the path ordered exponential formulation and collect the key kinematical identities: gauge covariance, composition under path concatenation, inversion, and base point dependence. We then explain how local information is encoded in suitable small segment and small loop limits, and how these limits enter reconstruction questions for the underlying gauge connection. Seen as an inverse problem, this perspective clarifies how much of the gauge connection is encoded in loop observables and why traced Wilson loops, while gauge invariant, are intrinsically lossy without additional input. Special attention is given to clarifying common sources of confusion, in particular the distinction between untraced holonomies and traced Wilson loops, and the role of reference path choices when relating open transports to based loop data. The presentation is intended to serve both as an entry point for newcomers and as a compact reference for readers in high energy theory, mathematical physics, and related fields.

Visual navigation requires goal directed planning informed by predictions of environment dynamics. Representative navigation world model methods propose complete action candidates before predicting consequences for selection or search updates. For each candidate, predictions primarily evaluate proposed actions rather than guide its stepwise construction. To this end, we propose InJepa, an intention first planning paradigm centered on latent states. A goal conditioned JEPA proposes latent intentions, and an action conditioned JEPA continually predicts future latent states. Both operate in a shared latent space to align their predictive representations. Within each candidate, predicted consequences feed into the next intention proposal and action inference, forming a closed loop of intention, action, and consequence. This extends the world model's role beyond evaluation to leading candidate construction through latent intentions. We evaluate InJepa through closed loop image goal navigation in Habitat MP3D. The results show effective multistep navigation through intention guided action generation and consequence feedback. Experiments across encoders further support applicability across visual representations. These findings support latent intentions as the central planning object and show that world models can directly guide plan construction through coordinated reasoning about intentions and consequences. Code is available at https://github.com/zhuxun132/InJepa.

Feedforward 3D reconstruction has largely followed an {appearance-driven paradigm}: multi-view images are encoded into visual representations, cross-view correspondence is implicitly inferred from learned features, and 3D primitives are subsequently predicted from these representations. However, calibrated cameras and sparse metric measurements already provide explicit constraints on where the same physical point should appear across views. We propose a novel paradigm, Geometry-Conditioned Feedforward Reconstruction, in which metric geometry establishes cross-view correspondence before the 3D representation is formed, while learned visual features refine rather than discover this correspondence. Based on this paradigm, we introduce GeoForward. Sparse LiDAR points are unprojected into 3D and analytically projected across calibrated cameras to provide metric correspondence anchors. A bounded continuous local search then refines these anchors using visual features, and the resulting geometry-conditioned representation is fused with global multi-view context from VGGT. The shared representation drives separate metric-depth and Gaussian-attribute prediction, establishing a direct path from explicit geometry to both Gaussian centers and spatial attributes. A sparse 3D refinement module subsequently improves local attribute compatibility while preserving the geometrically determined centers. For temporal reconstruction, endpoint Gaussians are composed using vehicle masks and motion priors. On nuScenes, GeoForward achieves the lowest cross-camera depth-consistency error and Chamfer distance among the compared methods while maintaining competitive reconstruction quality. These results demonstrate that explicitly conditioning correspondence formation on metric geometry provides an effective alternative to purely appearance-driven feedforward 3D reconstruction.

World Action Models (WAMs) jointly predict robot actions and future observations. However, their visual prediction objectives do not explicitly distinguish task-relevant transitions from incidental scene changes. For whole-body policies, this distinction matters because task progress depends on coordinated changes in body pose, contact, and object state. We introduce LANGUAGE-FORCING, a training strategy that uses language to emphasize task-relevant changes in the future visual representations of a whole-body WAM. Given paired current and ground- truth short-horizon future frames, a frozen vision-language model assesses task progress and predicts plausible subsequent state changes. We encode these descriptions with a frozen text encoder and align their embeddings with intermediate, action-conditioned future-visual features, complementing the original visual and action prediction objectives. All auxiliary components are used only during training, introducing no additional inference overhead. On HumanoidArena, LANGUAGE-FORCING achieves an average success rate of 79.17%, exceeding the strongest competing policy by 9.46 percentage points and the same WAM without language alignment by 10.77 percentage points. Across five real-world tasks, language alignment raises average success from 53.33% to 71.33% over the same WAM baseline.

Transformer compression often couples two distinct quantities: retained attention capacity and the number of sequential depths at which that capacity is executed. We study whether useful attention functions in a hierarchical vision transformer are locally mobile across depth. Using SegFormer-B5 for human segmentation, we find that some task-relevant attention heads can be relocated to nearby depths, but only when their computational context is respected. Head-only relocation is insufficient, while joint adaptation of the relocated head and surrounding representation boundary improves functional preservation; replacing donor-specific normalization and spatial-reduction preprocessing causes large degradation. We use this path-conditioned mobility to reorganize retained attention from depth into width. Starting from a fixed parent with 120 retained Stage-3 heads across 40 attention execution depths, we consolidate the same 120-head budget into 20 execution hubs. Under a layout-matched three-seed recovery control, K20 reaches 0.960686 +/- 0.000059 mIoU versus 0.960400 +/- 0.000074 for the corresponding State120 control. The paired point difference is +0.000285, while a paired image-bootstrap confidence interval includes zero; we therefore interpret the result as no measurable quality penalty rather than statistical superiority. For deployment, FOLDED_VEC preserves donor-specific paths while vectorizing their execution. Relative to the same 120-head State120 parent, it reduces end-to-end NVIDIA L4 latency by 5.28%, 3.15%, and 3.58% at batch sizes 1, 4, and 8. These results suggest that attention redundancy includes not only removable capacity but also execution-depth redundancy, provided that the computational path defining a relocated function is preserved.

Long video understanding requires reasoning over hours of content that far exceed the context window of multimodal large language models (MLLMs). Existing approaches compress video into fixed-length clips, yet events are fragmented and unrelated content is merged, yielding noisy retrieval. To overcome this challenge, we propose EMMA, an Event-driven Multimodal Memory system with Multi-Agent retrieval, which organizes memory around semantic events in three stages: (1) Event segmentation: a dual-stream detector adaptively locates event boundaries by fusing visual semantic drift with spoken transitions, rather than fixed windows. (2) Memory construction: built on the detected events, a three-level hierarchy of raw video, clip-level memory, and event-level memory stores content at progressively higher semantic abstraction. (3) Memory retrieval: a multi-agent framework collaboratively queries this hierarchy through query decomposition, tiered search, evidence validation, and answer generation, narrowing evidence from coarse events to precise clips. Extensive experiments on three challenging benchmarks demonstrate substantial improvements, showing the effectiveness of our event-aligned memory system. Code, prompt templates, and reproduction scripts are available at https://github.com/BAI-LAB/EMMA.

A robotic agent operating in an unknown environment must gather sufficient task-relevant information before acting. However, determining when the available information is sufficient remains a challenge. We propose a conceptual hybrid neuro-symbolic framework for task-sufficient active perception in unknown environments. After interpreting a requested task, the agent generates an initial plan and iteratively reevaluates it as new information becomes available. At each iteration, it can acquire a more informative viewpoint, inspect or manipulate candidate objects, ask the user for clarification, or perform the task. Neural perception modules update an object-centric belief state, while a symbolic decision layer determines whether the remaining uncertainty justifies further information gathering. The framework will be implemented on a mobile manipulator comprising a humanoid torso and a mobile base, with complementary evaluation in simulation. The proposed approach aims to enable robots to determine not only how to gather additional information, but also when further perception is no longer necessary.

Long-term personalized agents are required to reason over distributed history across modalities and evolving dialogue states. While existing flat cross-modal methods compress visual evidence prematurely and struggle with temporally stale observations, they neglect the necessity to unify multimodal and cross-session evidence. Crucially, raw multimodal memories are inherently entangled and heterogeneous, making fine-grained visual grounding and cross-modal reasoning computationally prohibitive. To this end, we present latent multimodal memory reasoning, i.e., LMMR, a novel neuro-symbolic framework that shifts long-term multimodal memory organization into a continuous latent space. Specifically, (i) instead of converting images to text and relying on static retrieval, LMMR maps both textual hidden states and visual patch features into a shared Sparse Autoencoder concept space, establishing sparse activations as common semantic addresses to dynamically construct a turn-level multimodal memory graph. (ii) A graph encoder then aggregates cross-modal and temporal dependencies across the task-specific latent subgraph, injecting the graph readout as a continuous residual rollout into the generation backbone while retaining native visual tokens to preserve fine-grained details. Extensive experiments on nine Mem-Gallery tasks and twelve MemEye settings demonstrate that LMMR significantly outperforms state-of-the-art baselines in capturing complex multimodal dependencies while enabling faithful and grounded responses.

Locally recoverable codes (LRCs) permit reconstruction of a lost coordinate from a small subset of the remaining coordinates, a property that is a fundamental coding technique for distributed storage and, more recently, for fault-tolerant quantum memories and distributed storage systems. Existing constructions of quantum locally recoverable codes (qLRCs) and their entanglement-assisted generalizations (EA-qLRCs) impose a single locality parameter $r$ governing recovery from both bit-flip ($X$-type) and phase-flip ($Z$-type) errors, even though physical noise is rarely symmetric between the two. We introduce entanglement-assisted quantum locally recoverable codes (EA-qLRC) with asymmetric locality $\{r_X,r_Z\}$, in which $X$-type and $Z$-type errors are corrected from recovery sets of size at most $r_X+1$ and $r_Z+1$, respectively. We extend a stabilizer sufficient condition for this property and, in particular, the CSS framework, where asymmetric locality corresponds to separate locality conditions on the two classical codes $\mathcal C_1$ and $\mathcal C_2$ underlying the construction. From this characterization we derive Singleton-, Griesmer-, Plotkin-, sphere-packing-, and Cadambe\textendash Mazumdar-like bounds for CSS-type asymmetric EA-qLRCs (AEA-qLRCs) built from Proposition~\ref{CSS-EA-qLRC-Asym}, and we characterize exactly when a pure CSS-type AEA-qLRC built from a pair of classical LRCs attains the Singleton-like bound with equality.

Multi-object grasping is a fundamental yet challenging problem, constrained by data collection costs, varying spatial layouts, and the need for broad generalization. To address these challenges, we propose MoDex, a data-efficient framework that provides the first unified solution for universal multi-object dexterous grasping across diverse spatial layouts. MoDex accomplishes this by formulating the task as a single-step decision problem of learning residual corrections to a reference trajectory, and introducing a spatially adaptive dual-strategy that performs single-shot grasping for proximal objects and sequential interaction for separated objects. Extensive evaluations show that MoDex significantly outperforms baselines in proximal configurations, achieving 61.24% success on unseen objects, and achieves high success rates in separated layouts. Further experiments demonstrate its generalization across robot embodiments, scalability and zero-shot sim-to-real deployment through a vision-based policy trained solely on simulation data. Code and pretrained models are available at https://github.com/Aurora-Xiong/MoDex.

Vision-language models are increasingly used for driving-scene understanding, yet their defensive reasoning under weakly structured Global South traffic remains unclear. We introduce BD-HazardVLM, a diagnostic benchmark of 500 Bangladesh road-scene images, comprising 375 from RSUD20K and 125 from TFP-BD, annotated for hazard presence, primary hazard category, risk timing, severity, defensive actions, visible evidence, and hallucination- and Western-road-bias traps. We evaluate Qwen2.5-VL-7B, InternVL3-8B, and MiniCPM-V-4.6 using a shared zero-shot task definition and structured output schema. Binary hazard-presence accuracy ranges from 88.0 percent to 89.2 percent, essentially matching the 89.0 percent majority-class baseline. Fine-grained reasoning is substantially weaker: category accuracy remains below 40 percent, risk-timing accuracy ranges from 1.6 percent to 12.4 percent, and exact defensive-action matching remains below 2 percent. Although InternVL3-8B achieves the highest severity accuracy at 61.2 percent, it has zero recall on the 58 high-severity cases. Qwen exhibits severe category mode collapse, while all models show weak latent-hazard anticipation and near-zero self-reported trap recognition. For the evaluated models, these findings show that coarse hazard recognition does not imply reliable fine-grained defensive-driving reasoning and motivate geographically diverse safety evaluation.

Buoyant plumes in stratified ambients arise in wildfire smoke transport, ventilation, and pollutant dispersion, where many-query prediction requires models that stay accurate and stable as the dynamics transition from free rise to internal-wave trapping. Operator Inference (OpInf) learns polynomial reduced models from snapshots, and the Boussinesq plume depends affinely on the squared buoyancy frequency $N^2$. However, this compatibility masks two stability gaps: a single affine operator family can distort the per-parameter dynamics into unstable spectra, and for quadratic OpInf models, finite-time blowup is governed by a cubic energy injection channel from the inferred quadratic operator. To address these gaps, we develop cyclic-skew regime-bridged OpInf (CR-OpInf). We derive an affine-inadequacy diagnostic that links the departure from a shared affine family to destabilizing spectral drift before integration. In place of the shared affine fit, we interpolate per-parameter operators with a Gaussian kernel. We also extend the cyclic-skew quadratic constraint to the buoyancy-source-forced parametric setting, closing the cubic energy channel and isolating this closure as the boundedness mechanism. We further introduce a regime-bridging basis split that resolves free rise and trapped intrusion structures while preserving the cyclic-skew structure under convex interpolation. Applied to an unconfined Boussinesq plume and a viscous Burgers problem spanning a smooth-to-shock transition, CR-OpInf remains bounded for five training horizons, avoids the divergence of affine and unconstrained variants, tracks the plume rise, and recovers stratification-dependent rise scaling at moderate horizons. To our knowledge, this is the first parametric OpInf study of an unconfined stratified plume.

We study the problem of learning diffusion-based world models that jointly capture scene geometry and visual appearance in complex first-person environments. Existing approaches typically train a single action-conditioned model that predicts future RGB frames, but they either ignore geometry or require action supervision for all modalities, limiting scalability to large unlabeled video datasets. In this work, we introduce a two-stage diffusion world model that separates geometric and appearance prediction while maintaining joint temporal coherence. Our key idea is a staggered RGB-depth formulation: depth at time $t+1$ is predicted first using an action-conditioned diffusion transformer, and the RGB frame at the same timestep is then generated conditioned on this predicted depth using a second diffusion model without action inputs. We show that this decoupled design faithfully approximates a hypothetical joint RGBD diffusion model trained with diffusion forcing, enabling the second stage to be trained or fine-tuned on additional RGB-depth data without requiring action labels. Experiments on the CS:GO Dust 2 dataset demonstrate that our approach improves both depth accuracy and RGB fidelity compared to fully joint models and RGB-only baselines, while significantly enhancing scalability to augmented or unlabeled datasets. Our results highlight the effectiveness of geometry-guided diffusion for world modeling and open a path toward large-scale action-free video model training. Code to be released upon acceptance of paper.

Damage prediction in particulate elastomeric composites is challenging due to the strong material heterogeneity, nearly incompressible behavior of the matrix, and the presence of localized stress concentrations arising from matrix-inclusion interactions. In the present work, a computational framework is developed for damage propagation analysis of particulate elastomeric composite structures. The effective constitutive response of the heterogeneous material is first determined using the second-order homogenization method. The resulting homogenized response is subsequently used to identify equivalent hyperelastic material parameters for structural simulations. A three-parameter isotropic damage model is incorporated to capture damage-induced softening. A mixed displacement-pressure finite element formulation, is used to alleviate volumetric locking associated with nearly incompressible property of elastomers. The proposed framework is validated through full-scale simulations and homogenized response involving composites with varying inclusion volume fractions. Numerical investigations demonstrate that the homogenized models is able to reproduce the global force-displacement response within 2% error and damage evolution characteristics of the heterogeneous composites while significantly reducing the computational cost.

This study investigates generalized boundary representations and transfer learning for physics-informed neural-network (PINN) estimation of spatially varying convective heat transfer coefficients (HTCs) in jet impingement cooling. Piecewise-linear (PWL) and neural-network boundary estimators are inferred from transient interior-temperature measurements and evaluated against CFD-CHT reference HTC profiles at Reynolds numbers of 5000, 10000 and 15000. A discrete Adjoint-FEM formulation with gradient regularization based on the $H^1$-seminorm provides an independent inverse-method comparison. Transfer learning is investigated under two source-target conditions: transfer between noise levels at the same Reynolds number, and transfer from a controlled Gaussian-profile problem to CFD-CHT-derived target problems at different Reynolds numbers. For the representative Re = 5000 target with 5% measurement noise, quadratic, PWL, and neural-network PINNs achieve relative HTC-profile errors of 5.08, 5.31 and 5.19%, respectively, while generalized representations reduce the error for the Re = 15000 target from 16.76% for the quadratic PINN to 14.41% and 14.46%. Transfer learning reduces target-training time by 57.86-81.71% for fixed-Reynolds-number transfer and by 50.59-84.94% for Gaussian-to-CHT transfer. Its effect on reconstruction accuracy is target dependent: fixed-profile PWL transfer remains within 0.30 percentage points of scratch training, whereas Gaussian-to-CHT transfer produces differences of 0.10-1.52 percentage points across targets and representations. These results demonstrate that generalized PINN representations can accommodate spatially varying HTC profiles without a prescribed quadratic form, while transfer learning can substantially reduce target-training cost for repeated inverse estimation, with the resulting reconstruction accuracy depending on source-target compatibility.

A familiar line of reasoning runs as follows: language models are flexible approximators trained under a strictly proper loss, so scale brings them closer to Solomonoff induction, and their predictive success is evidence of that approach. We apply known results from algorithmic information theory, universal prediction and Bayesian consistency. In its uniform form the conclusion holds for no computable model: each has cheap environments it never learns, and every uniformly computable family of total predictors shares one on which each member fails beyond a horizon growing with its index, so “closer” can hold only pointwise, one environment at a time. Nor could the offered routes certify even that: finite evaluation cannot certify agreement on rarely reached contexts, perfect corpus fit selects the corpus law as prior, and asymptotic criteria cannot see the constants that carry Solomonoff's guarantee. Our lens is the Kraft inequality read for predictors. Every model has at most the same budget of predictive success, so Solomonoff's predictor, which leaves part of it unspent, differs from a language model not in having more of it but in where it spends it; on mutually singular families this is exact, and only full-mass Bayes mixtures reach the frontier. We then discuss what is inheritable, including in-context learning as amortised Bayesian inference; what survives within a context window; which Solomonoff-like properties a deployed system should have or avoid; and what served conditionals determine. A closing table states what common practices certify and what to report. Tests are developed in a companion paper.