arXiv:2605.19137v1 Announce Type: new
Abstract: Video foundation models achieve strong performance across many video understanding tasks, but typically require large-scale pre-training on massive video datasets, resulting in substantial data and compute costs. In contrast, modern image foundation models already provide powerful spatial representations. This raises an important question: can competitive video models be built by reusing these spatial representations and pre-training only for temporal reasoning? We take initial steps toward exploring a lightweight training paradigm that freezes a pre-trained image foundation model and trains only a recurrent temporal module to process streaming video. By reusing an image foundation model as a spatial encoder, this approach could significantly reduce the amount of video data and compute required compared to end-to-end video pre-training. In this work, we explore the feasibility of this approach before investing in computing for video pre-training. Our empirical findings across multiple video understanding tasks suggest that strong temporal performance can emerge without large-scale video pre-training, motivating future work on recurrent video foundation models obtained by pre-training a temporal module on top of a frozen image foundation model. Code: https://github.com/tue-mps/towards-video-image-frozen .
Science Journals
arXiv:2605.19135v1 Announce Type: new
Abstract: Causal representation learning (CRL) seeks to uncover meaningful latent variables and their corresponding causal structure from high-dimensional observational data. Although its significance, CRL identifiability remains a crucial property, as it ensures the recovery of the mechanisms behind the data generation process, and hence the interpretability and robustness of the representation. Proving identifiability in CRL is intrinsically difficult, and we address in this work an even more challenging setting: multimodality. We consider multimodal observed data with a latent partially shared structure. Each modality is generated, through non linear mixing functions, from a specific subset of causal latent variables. Under flexible assumptions and without imposing any parametric distribution on the latent variables, we establish component-wise identifiability guarantees for the causal latent representation. Our identifiability results, furthermore, apply to the undercomplete scenario where we have, for each modality, more observed than latent variables. To instantiate our theoretical analysis, we introduce a Wasserstein-based module to recover the partially shared latent structure. Due to its differentiability, the latter can be easily integrated into all types of architecture, only requiring minimal changes. Extensive experiments on synthetic and realistic datasets validate the superiority of our approach over SOTA methods.
arXiv:2605.19306v1 Announce Type: new
Abstract: We address the open problem of training hypernetworks for Controllable Pareto Front Learning (CPFL) under split feasibility conditions with rigorous theoretical guarantees. We reformulate the constrained Pareto problem as a Bi-Level Scalarized Split Problem (BSSP) and propose the Adaptive Balanced Penalty (ABP) algorithm, whose three gradient components -- optimality, set feasibility, and image feasibility -- are blended through an adaptive indicator driven by a computable lower bound. Using a novel convex surrogate technique, we prove full-sequence convergence under standard convexity and Robbins-Monro step-size assumptions. The ABP penalty structure is then translated into a two-phase, feasibility-first training strategy for Hyper-MLP and HyperTrans architectures (ABP-HyperNet). To evaluate constrained CPFL, we introduce the Expected Feasible Hypervolume (EFHV), which jointly captures solution quality and constraint satisfaction. Experiments on five multi-objective benchmarks validate the ABP solver against ground truth, while three multi-task learning datasets demonstrate that ABP-HyperNet achieves up to 2.3x higher EFHV than unconstrained baselines by raising feasibility from 36-49% to 87-100%.
arXiv:2605.20151v1 Announce Type: new
Abstract: The proliferation of generative artificial intelligence has given rise to an interactive learning environment, where model parameters are continuously updated using not only data generated by natural processes, but also synthetic outputs produced by other models. This paradigm introduces two major challenges: (1) training data are no longer drawn exclusively from the target population, undermining a core assumption of classical statistical learning, and (2) model training processes become inherently correlated, as models interact with one another through repeated exposure to each other's synthetic outputs in a potentially complex manner. Establishing reliable statistical inference in such structured interactive learning environments therefore remains an important open problem. In particular, there is growing concern about model collapse, a phenomenon in which the performance of generative models progressively degrades as they are trained on synthetic data produced by earlier model generations.
Prior work on model collapse primarily focuses on a single model trained on its own output, failing to capture model performance in multi-model interactive settings. In this work, we fill this gap by investigating the performance of generative models in an interactive learning environment with general interaction patterns. In particular, we formalize model interactions using directed graphs and show that the occurrence of model collapse depends critically on the topology of the interaction graph. We further derive an explicit necessary and sufficient condition characterizing when model collapse occurs, and establish finite-sample results for linear regression and asymptotic guarantees for general M-estimators. We support our theoretical findings through extensive numerical experiments.
arXiv:2605.19118v1 Announce Type: new
Abstract: On January 1, 2024, a moment magnitude (Mw) 7.6 earthquake struck the Noto Peninsula, Japan, causing intense ground shaking and triggering a tsunami along the Japan Sea coast. Preliminary analysis by the Japan Meteorological Agency (JMA) identified a reverse-fault rupture consistent with a northwest-southeast compressional stress regime. Aftershock distribution analysis (JMA, 2024) revealed that the causative fault extended approximately 150 km from the western Noto Peninsula to the northeastern offshore area, aligning with the inferred tsunami source region. While the rupture mechanism and impacts have been studied, high-resolution seismic imaging of the shallow crustal structure within the rupture zone remains limited. To address this gap, the Atmosphere and Ocean Research Institute (AORI) at the University of Tokyo conducted a multichannel seismic (MCS) reflection survey aboard the R/V Hakuho-Maru in March 2024, collecting high-quality MCS data along 14 profiles (approx. 45 km each). The data were processed using an advanced depth imaging workflow incorporating grid-based tomography refined by automated continuity attributes to enhance reflection coherency. Structural attributes (dip and continuity) were extracted from migrated sections and used for automated horizon picking via seismic pencil construction. The P-wave velocity model was iteratively refined using grid-based tomography to optimize horizon alignment and minimize residual moveout (RMO) in migrated common image gathers. The resulting 2D seismic sections and 3D visualizations provide the first high-resolution images of the shallow rupture zone associated with the 2024 Noto earthquake. This dataset offers a critical foundation for ongoing research into fault geometry, rupture dynamics, and the broader seismotectonic framework of the region.
arXiv:2603.11024v3 Announce Type: replace
Abstract: VLMs have become increasingly proficient at a range of computer vision tasks, such as visual question answering and object detection. This includes increasingly strong capabilities in the domain of art, from analyzing artwork to generation of art. In an interdisciplinary collaboration between computer scientists and art historians, we characterize the mechanisms underlying VLMs' ability to predict artistic style and assess the extent to which they align with the criteria art historians use to reason about artistic style. We employ a latent-space decomposition approach to identify concepts that drive art style prediction and conduct quantitative evaluations, causal analysis and assessment by art historians. Our findings indicate that 73% of the extracted concepts are judged by art historians to exhibit a coherent and semantically meaningful visual feature and 90% of concepts used to predict style of a given artwork were judged relevant. In cases where an irrelevant concept was used to successfully predict style, art historians identified possible reasons for its success; for example, the model might "understand" a concept in more formal terms, such as dark/light contrasts.
arXiv:2605.19316v1 Announce Type: new
Abstract: Recent studies in difficulty-controlled reading comprehension item generation have leveraged large language models (LLMs) to produce items by adjusting difficulty-related features. However, existing methods typically rely on a single-agent prompting approach, which often fails to consistently satisfy specified feature constraints, resulting in items that deviate from the target difficulty level. To address this limitation, we introduce MAFIG, a Multi-agent Framework for Feature-constrained Item Generation, where multiple LLM agents and feature-specific evaluators collaborate to generate and iteratively revise items based on intended constraints. Furthermore, to verify the efficacy of MAFIG in difficulty control, we propose a method for constructing a sequence of feature constraint sets that yield items with monotonically increasing difficulty. Experimental results demonstrate that MAFIG generates items that adhere to target constraints at a significantly higher rate than baselines, achieving robust difficulty control through the difficulty-calibrated constraint sequence.
arXiv:2605.19018v1 Announce Type: new
Abstract: Fine-tuning adapts a pre-trained model to downstream tasks using a small amount of labeled data. Low-Rank Adaptation (LoRA) is an efficient fine-tuning method that reduces memory and computation costs while often achieving performance close to full fine-tuning. Despite its widespread use, the theoretical behavior of LoRA is not yet well understood. In this paper, we study LoRA in a simple linear regression setting and compare its excess risk with that of full fine-tuning. Our analysis identifies regimes in which LoRA achieves lower excess risk than full fine-tuning in both overdetermined and underdetermined settings. Specifically, our theory predicts that LoRA can outperform full fine-tuning when the difference between the pretraining and the downstream tasks is effectively low-rank. We further show how the choice of LoRA rank affects generalization performance, explaining why using a very small rank can improve test accuracy in certain settings, even though it limits model expressivity. Finally, we support our theoretical results with experiments on practical tasks, suggesting that the identified tradeoffs and insights extend beyond linear regression.
arXiv:2605.18998v1 Announce Type: new
Abstract: Recent experiments have shown contradictory effects of strong light-matter coupling on exciton-exciton annihilation (EEA) in organic molecular systems. In this work, we perform numerical simulations of polariton dynamics and reveal the role of strong coupling in changing the EEA rate. The results of our simulations suggest that strong coupling allows to partially overcome disorder in systems with poor exciton mobility via delocalisation of excitons owing to the interaction with the common cavity mode. This leads to an enhanced connectivity between excitons and, consequently, to an increase in the EEA rate. Conversely, in systems with high exciton mobility, in which disorder has a much smaller effect on excitation energy transfer, excitons can interact strongly even without coupling to the cavity photons at the exciton densities at which EEA typically occurs. In this case, the EEA rate can be even lower than in bare molecules due to the existence of a competing decay channel associated with photon leakage through the cavity mirrors. We also find that in the weak coupling regime, the EEA rate appears to be suppressed due to this decay channel regardless of the exciton transport properties. Our simulations resolve the experimental controversy on the effect of strong coupling on EEA and provide guidance for minimising the EEA rate towards a more feasible realisation of Bose-Einstein condensation of polaritons.
arXiv:2605.19410v1 Announce Type: new
Abstract: Segmentation has become easy when the concept is known, requiring retrieval of a learned visual grounding from text. It remains hard for open ad-hoc concepts, where the grounding may not exist as one learned mask and must often be constructed from image evidence through parts, relations, exclusions, and collections. We propose a Vision-guided Ad-hoc Segmentation Agent (VASA), the first vision harnessing agent for open ad-hoc segmentation. VASA is training-free and couples a VLM agent, a segmentation foundation model, and a visually grounded workflow. Rather than revising text prompts alone, VASA uses a persistent working mask to reason, construct, and validate a solution. It plans visual operations, invokes segmentation tools, inspects results, edits the mask, and recovers from errors. We construct PARS, a new benchmark that turns part-level labels in PartImageNet into open ad-hoc concepts through long-form definition queries. On PARS, VASA outperforms open-vocabulary, reasoning-based, and agentic baselines, surpassing SAM3 Agent by 14-25%. On RefCOCOm, a standard multi-granularity referring segmentation benchmark, VASA improves over SAM3 Agent by 5-9% and over other agentic baselines by up to 20%. These results validate agentic visual construction for open ad-hoc segmentation. Our work points to a path for AI agents beyond wrapping foundation models as tools: Programming them with task knowledge, VLM behavior, visual routines, working memory, and failure-aware workflows.
arXiv:2602.05709v2 Announce Type: replace
Abstract: Low-rank adaptation (LoRA) approximates the update of a pretrained weight matrix using the product of two low-rank matrices. However, standard LoRA follows an explicit-rank paradigm, where increasing model capacity requires adding more rows or columns (i.e., basis vectors) to the low-rank matrices, leading to substantial parameter growth. In this paper, we find that these basis vectors exhibit significant parameter redundancy and can be compactly represented by lightweight nonlinear functions. Therefore, we propose Generative Low-Rank Adapter (GenLoRA), which replaces explicit basis vector storage with nonlinear basis vector generation. Specifically, GenLoRA maintains a latent vector for each low-rank matrix and employs a set of lightweight radial basis functions (RBFs) to synthesize the basis vectors. Each RBF requires far fewer parameters than an explicit basis vector, enabling higher parameter efficiency in GenLoRA. Extensive experiments across multiple datasets and architectures show that GenLoRA attains higher effective LoRA ranks under smaller parameter budgets, resulting in superior fine-tuning performance. The code is available at https://anonymous.4open.science/r/GenLoRA.
arXiv:2511.13588v2 Announce Type: replace
Abstract: Model Predictive Control (MPC) is a powerful framework for optimal control but can be too slow for low-latency applications. We present a data-driven framework to accelerate MPC by replacing online optimization with a nonparametric policy constructed from offline MPC solutions. Our policy is greedy with respect to a constructed upper bound on the optimal cost-to-go, and can be implemented as a nonparametric lookup rule that is orders of magnitude faster than solving MPC online. Our analysis shows that under sufficient coverage conditions of the offline data, the policy is recursively feasible and admits provable, bounded optimality gap. These conditions establish an explicit trade-off between the amount of data collected and the tightness of the bounds. New solutions can be incorporated straightforwardly without the need for retraining, enabling continual improvement. Our experiments show that this policy is between 100 and 1000 times faster than standard MPC with only a modest hit to optimality, showing potential for real-time control tasks.
arXiv:2511.06943v3 Announce Type: replace
Abstract: Global plant maps of plant traits, such as leaf nitrogen or plant height, are essential for understanding ecosystem processes, including the carbon and energy cycles of the Earth system. However, existing trait maps remain limited by the high cost and sparse geographic coverage of field-based measurements. Citizen science initiatives offer a largely untapped resource to overcome these limitations, with over 50 million geotagged plant photographs worldwide capturing valuable visual information on plant morphology and physiology. In this study, we introduce PlantTraitNet, a multi-modal, multi-task uncertainty-aware deep learning framework that predictsfour key plant traits (plant height, leaf area, specific leaf area, and nitrogen content) from citizen science photos using weak supervision. By aggregating individual trait predictions across space, we generate global maps of trait distributions. We validate these maps against independent vegetation survey data (sPlotOpen) and benchmark them against leading global trait products. Our results show that PlantTraitNet consistently outperforms existing trait maps across all evaluated traits, demonstrating that citizen science imagery, when integrated with computer vision and geospatial AI, enables not only scalable but also more accurate global trait mapping. This approach offers a powerful new pathway for ecological research and Earth system modeling.
arXiv:2605.19340v1 Announce Type: new
Abstract: Vision foundation models (VFMs) have achieved strong performance across various vision tasks. However, it still remains challenging to apply VFMs for cross-domain few-shot segmentation (CD-FSS), which segments objects of novel classes under domain shifts using only a few labeled exemplars. The challenge is mainly driven by two factors: (1) limited labeled exemplars per novel class relative to the scale of VFM pre-training, making the model prone to overfitting during retraining, and (2) target-domain shifts underrepresented during pre-training, inducing cross-domain inconsistency and layer-wise sensitivity. To address these issues, we propose Hierarchical Exemplar Representation Adaptation (HERA), a three-stage select-regularize-calibrate VFM-based segmentation framework that learns effectively from limited labels and adapts to novel domains without source-data retraining. We first design Hierarchical Layer Selection (HLS) to adaptively identify the most informative VFM layer using a data-dependent Exemplar Transfer Risk (ETR) computed for each candidate layer. Then, Prior-Guided Regularization (PGR) regularizes interactions on the selected representation, yielding well-structured local signals for the subsequent stage. Furthermore, Pixelwise Adaptive Calibration (PAC) combines the selected representation with the refined interaction maps to calibrate pixel-wise predictions, producing consistent masks. Together, these stages form a hierarchical select-regularize-calibrate pipeline that guides frozen VFM features in new domains while fine-tuning less than 2.7% of parameters at test time. Extensive experiments show that HERA surpasses the state of the art by more than 4.1 mIoU across multiple CD-FSS benchmarks.
arXiv:2602.09023v4 Announce Type: replace
Abstract: Despite strong generalization capabilities, Vision-Language-Action (VLA) models remain constrained by the high cost of expert demonstrations and limited real-world interaction. While online reinforcement learning (RL) has shown promise, its application to real-world VLA manipulation is hindered by low exploration efficiency and restricted exploration coverage. Through systematic real-world experiments, we observe that the effective exploration space of online RL is largely constrained by the trajectory distribution induced during supervised fine-tuning (SFT). Motivated by this observation, we propose TwinRL, a digital twin-real-world collaborative post-training framework that expands and guides RL exploration for VLA models through three stages: SFT warm-up, twin RL warm-up, and real-world RL. TwinRL first reconstructs a high-fidelity digital twin from smartphone-captured scenes. During the SFT stage, we introduce an exploration space expansion strategy that expands the support of the trajectory distribution beyond real demonstrations, reshaping the exploration space for more effective RL. Rather than treating the twin as a data augmentation tool, we propose a twin RL warm-up strategy that enables it to act as an exploration guide for real-world RL. Specifically, TwinRL performs efficient parallel RL in the digital twin to generate interactive trajectories that populate the replay buffer and stabilize subsequent real-world RL learning. This process also identifies failure-prone yet informative configurations, enabling targeted human-in-the-loop rollouts to further improve on-robot efficiency. Across four tasks, TwinRL achieves near-100% success in both in-distribution and out-of-distribution regions, delivering over 30% faster convergence than prior real-world RL methods with only 20 minutes of on-robot interaction.
arXiv:2510.25897v2 Announce Type: replace
Abstract: The default paradigm of post-training text-to-image generators includes post-hoc selection of generated images, and subsequent training with one reward model to align the generator to the reward, typically user preference. This discards informative data as well as optimizes only for a single reward, hence harming diversity, semantic fidelity and efficiency. Instead, we propose MIRO, a method that conditions the model on multiple rewards during training, thus letting the model learn user preferences directly. MIRO pre-training both improves the visual quality of the generated images and speeds up the training, achieving state of the art on the GenEval compositional benchmark and user-preference scores (PickAScore, ImageReward, HPSv2).
arXiv:2605.19373v1 Announce Type: new
Abstract: All 26 neural network merge strategies we tested including weight averaging, SLERP, TIES, DARE, Fisher merging, and evolutionary approaches -- fail the algebraic properties (commutativity, associativity, idempotency) required for conflict-free distributed operation. We prove that this failure is structural: normalisation-based merges cannot simultaneously satisfy all three properties. To resolve this, we present a two-layer architecture -- CRDTMergeState -- that wraps any merge strategy in a CRDT-compliant (Conflict-Free Replicated Data Type) layer. Layer 1 manages contributions via OR-Set CRDT semantics, where the merge operation is set union -- trivially commutative, associative, and idempotent. Layer 2 applies merge strategies as deterministic pure functions over a canonically-ordered contribution set, with randomness seeded from the Merkle root. We prove that this separation guarantees Strong Eventual Consistency: all replicas receiving the same contributions compute identical merged models, regardless of message ordering. Empirical validation spans three tiers: controlled 4x4 tensors (104/104 tests pass), production-scale models up to 7.24B parameters (208 strategy-level tests, 43,368 layer-level property checks at capped tensor resolution), and multi-node convergence under gossip and partition healing (100 nodes, 20 orderings), with CRDT overhead below 0.5 ms. Because the wrapper is transparent, downstream performance is identical by construction, confirmed via byte-identical output verification. The reference implementation is available as crdt-merge v0.9.4.
arXiv:2605.19380v1 Announce Type: new
Abstract: This letter presents a comprehensive analysis of the stability phenomenon related to the ability of generators to remain in synchronism when subjected to small or large disturbances, in power systems with both synchronous machines and grid-forming voltage source converters (GFM-VSC). This phenomenon is associated with two stability classes in the IEEE/PES classification, namely, rotor-angle stability (when involving synchronous machines and slow-interaction converter-driven stability (when involving power converters). However, this work shows that this phenomenon is fully characterised with the slow dynamics of the angle difference between the voltage sources connected to the power system, regardless of whether they are synchronous machines (with rotors) or GFM-VSCs. Therefore, we suggest using the term angle stability to refer to this phenomenon, while slow-interaction converter-driven stability should only include slow interactions of different nature involving power converters.
arXiv:2605.19382v1 Announce Type: new
Abstract: Programmatic video generation through code offers geometric precision and temporal coherence beyond pixel-level diffusion models, yet rigorously evaluating whether language models can produce spatially correct animated outputs remains an open problem. We introduce PRISM, a large-scale benchmark of 10,372 human-calibrated instruction-code pairs (20 times larger than prior programmatic video generation benchmarks), grounded in real-world knowledge visualization scenarios across English and Chinese and spanning 437 subject categories. We further propose a funnel-style evaluation framework with four complementary metrics: Code-Level Reliability for executability, Spatial Reasoning for layout correctness over full animation sequences, and Prompt-Aware Dynamic Visual Complexity (PADVC) and Temporal Density (TD) for diagnosing dynamic expression and temporal activity. Systematic evaluation of seven mainstream LLMs reveals a striking Execution-Spatial Gap: the average drop from execution success rate to spatial pass rate is approximately 41%, showing that runnable code does not necessarily yield spatially coherent visual output. These findings show that programmatic video generation evaluation should go beyond executability. PRISM provides a principled benchmark for advancing spatially coherent code generation.
arXiv:2602.13076v2 Announce Type: replace
Abstract: Surfactants reside at the interface of two-phase flows and significantly influence the flow dynamics. Numerical simulations are essential for a comprehensive understanding of such surfactant-laden flows and require a method that can accurately simulate surfactant transport along the interface. In this study, we focus on interfacial transport models for insoluble surfactants based on the diffuse-interface method and propose two approaches to improve their accuracy: (a) adopting a formulation that avoids the spatial derivatives of variables with sharp gradients and (b) allowing the width of the delta function to be specified independently of the interface width. These approaches are simple and practical in that they do not lead to significant increases in computational cost, implementation complexity, or degradation of interface-capturing accuracy. Moreover, they preserve the discrete conservation of both fluid and surfactant mass. We conduct a series of numerical tests to demonstrate the effectiveness of the proposed approaches. Finally, we present a challenging test case that is difficult to solve accurately and has not been previously discussed. We expect this case to serve as a valuable benchmark for evaluating and comparing the performances of various methods proposed in the literature.
arXiv:2602.23622v2 Announce Type: replace
Abstract: Significant progress has been made in the field of Instruction-based Image Editing Models (IIEMs). However, while these models demonstrate plausible adherence to instructions and strong reasoning ability on current benchmarks, their ability to edit small objects remains underexplored, despite its importance for precise local editing and refining details in both real and generated images. In this paper, we introduce DeepLookEditBench (DLEBench), the first benchmark dedicated to assessing the abilities of IIEMs in editing small-scale objects. Specifically, we construct a challenging testbed comprising 1889 samples across seven instruction types. In these samples, target objects occupy only 1%-10% of the image area, covering complex scenarios such as partial occlusion and multi-object editing. To ensure robust evaluation on this benchmark, we propose an evaluation protocol with refined score rubrics to minimize subjectivity and ambiguity in two criteria: Instruction Following and Visual Consistency. This protocol also introduces a dual-mode evaluation framework (Tool-driven and Oracle-guided Modes) addressing the misalignment between LMM-as-a-Judge and human judgements on DLEBench. Empirical results on 10 IIEMs reveal significant performance gaps in small-scale object editing, highlighting the need for specialized benchmarks to advance this ability.
arXiv:2605.19557v1 Announce Type: cross
Abstract: We study post-hoc Learning to Defer (L2D) through the lens of ideal distributions: divergence-regularized reweightings of the data distribution under which a model attains low loss. We define deferral via the density-ratio between a model's and an expert's ideals. Using the reduction from density-ratio estimation to class-probability estimation, we derive the DR CPE losses for post-hoc L2D scorers. Deferral decisions are then made by thresholding the scorer, allowing deferral rates to be adjusted without retraining. For KL-based ideal distributions, our deferral rules recovers Chow's rule under the original distribution and a connection to an expert-tilted Bayes posterior -- which incorporates the expert's performance -- depending on if the ideal distributions are joint or marginal distributions. Experimentally, our approach is competitive compared to common baselines and more robust across dataset settings. More broadly, our results cast post-hoc L2D as density-ratio learning between ideal distributions, bridging Chow-style rules, expert comparison, and elucidating connections to related learning settings including anomaly detection.
Operational Memory Architecture for Kubernetes:Preserving Causal Context Across the Evidence Horizon
arXiv:2605.18755v1 Announce Type: new
Abstract: Kubernetes clusters generate rich operational events during pod lifecycle transitions, yet the platform's native event retention model discards the most diagnostically valuable context. The LastTerminationState field, which records a container's last failure, is overwritten shortly after a pod restart. We define this as the evidence horizon. During high-frequency crash loops, this horizon may be crossed multiple times before inspection, permanently losing critical evidence.
This paper introduces the Operational Memory Architecture (OMA) to preserve causal failure evidence before event rotation. OMA encodes evidence retention and causal reconstruction as explicit architectural requirements. It captures operational events into causal chains using three patterns: P001 (OOMKill chain), P002 (ConfigMap variable misconfiguration), and P003 (ConfigMap volume mount propagation).
We implement OMA as an open-source system with a Go-based Kubernetes watcher, SQLite operational memory store, and a simple query interface. Experiments on Minikube and AKS include a 30-run latency analysis and stress tests with up to 20 crash-looping pods. Causal edges are built with mean latency below 1 ms. The collector processes ~2.8 events/sec while using under 10 MB memory, showing minimal overhead and effective evidence preservation.
arXiv:2602.23788v2 Announce Type: replace
Abstract: The Satellite Internet of Things (S-IoT) enables global connectivity for remote sensing devices that must operate energy-efficiently over long time spans. We consider an S-IoT system consisting of a sender-receiver pair connected by a data channel and a feedback channel and capture its dynamics using a Markov Decision Process (MDP). To extend battery life, the sender has to decide on deep-sleep durations. Deep-sleep scheduling is the primary lever to reduce energy consumption, since sleeping devices consume only a fraction of their idle power. By choosing its deep-sleep duration online, the sender has to find a trade-off between energy consumption and data quality degradation at the receiver, captured by a weighted sum of costs. We quantify data quality degradation via the recently introduced Goal-Oriented Tensor (GoT) metric, which can take both age and content of delivered data into account. We assume a Markovian observed process and Markov channels with time-varying delay and erasure rates. The challenge is that content awareness of the GoT metric makes periodic transmissions inherently inefficient. Additionally, optimal sleep durations depends on the (unknown) future states of the observed process and the channels, both of which must be inferred online. We propose a novel algorithm using probabilistic simulation-based optimization (PSBO). With PSBO, the sensor forecasts future states based on estimated transition probabilities, and uses these forecasts to select the optimal deep-sleep duration. Extensive simulations and experiments with S-IoT hardware demonstrate superior performance of PSBO under diverse conditions.
arXiv:2603.00995v2 Announce Type: replace
Abstract: The rapid progress in 6G communication and high-bandwidth radar has driven an unprecedented surge in the spatial density of signal sources, resulting in an increasingly congested electromagnetic (EM) environment. When resolving closely spaced signals and interference, existing architectures are strictly bounded by the inherent diffraction limits of two-dimensional (2D) physical apertures, hindering super-resolution sensing and multi-interference mitigation in complex scenarios. Here, we present a 3D aperture-engineered diffractive neural network (AE-DNN) that achieves super-resolution sensing and computing by extending the traditional 2D aperture into 3D. The 3D aperture engineering framework is realized by constructing deep cascaded metasurface layers so that the diffractive propagation from oblique incident fields can be layer-wise modulated and piecewise encoded for perceiving EM fields far exceeding physical aperture limits. The N-layer AE-DNN has the capability to achieve ~N times higher angular resolution than the 2D aperture diffraction limit. The multi-dimensional synthetic aperture (MSA) training is developed to achieve speed-of-light coherent synthesis of the 3D aperture and integrate neural network-based modeling of multi-dimensional metasurface modulation. By orthogonalizing array response vectors in the analog domain, AE-DNN performs parallel super-resolution angle estimation, source number estimation, and source separation for up to 10 independent coherent or incoherent sources. Experimental results across the 36-41 GHz band demonstrate that AE-DNN resolves and suppresses closely spaced multi-interference by ~20 dB, enhances communication capacity by 13.5X, and reduces latency by three orders of magnitude. AE-DNN heralds a paradigm shift in signal processing for advanced radar and 6G communications.