# Research Papers — 2026-09 · bundle page 234 of 266 > Full-text markdown of papers #11651–11700 of 13298 published in 2026-09, newest first. AI-readable mirrors of the full 13298-paper 2026-09 archive (50 papers per file). Canonical page: https://aipod.dev/papers/archive/2026-09/page/234 [← Previous bundle](https://aipod.dev/papers/archive/2026-09/page/233.md) [Next bundle →](https://aipod.dev/papers/archive/2026-09/page/235.md) ### Amortizing Scaling Law Construction Costs Authors: Abhash Kumar Jha, Diana Alexandra Onuţu, Neeratyoy Mallik, Swagatam Haldar, Sam Laing, Niccolò Ajroldi, Shiwei Liu, Joaquin Vanschoren, Aaron Klein Published: 2026-09-04 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2609.05016v1 Scaling laws guide the design choices for training large foundation models, but deriving them involves training an exhaustive grid over hyperparameters, token budgets, and parameter counts, which is computationally expensive. Fitting a scaling law, however, only requires the best-loss frontier across compute scales, discarding most of the trained configurations. We propose a framework for efficient scaling law construction that formulates data collection as a Bayesian optimization problem, and introduce metrics for comparing scaling law fitting methods under constrained compute budgets. We find that progressively expanding the compute budget during acquisition, mirroring the compute-ordered evaluation of configurations in practice, substantially improves recovery efficiency. Augmenting the observed configurations with surrogate-fantasized evaluations then recovers the broader experimental grid, allowing accurate scaling law fitting without training every configuration. Together, these can closely match scaling law fits over a full dense grid at computational savings of up to $10\text{--}100\times$. --- ### Solution-space heterogeneity shapes federated learning dynamics across partial differential equations Authors: Ping Luo, Jiahuan Wang, Ziqing Wen, Tao Sun, Dongsheng Li Published: 2026-09-04 Categories: cs.LG, cs.DC arXiv: https://arxiv.org/abs/2609.05012v1 Federated scientific machine learning enables institutions to train neural surrogates without centralizing local physical data, yet studies of partial differential equations (PDEs) lack a transferable definition of non-independent and identically distributed data. Existing protocols partition coordinates, coefficients, boundary conditions, or geometries according to equation-specific rules. Here, we introduce solution-space PDE-Dirichlet, a protocol that converts continuous supervised responses into reusable solution bins and quantifies the realized separation between clients through optimal transport over the geometry of these bins. We derive an exact inverse relation between population allocation heterogeneity and the Dirichlet concentration, and we establish conditions under which response heterogeneity induces gradient disagreement, local-update dispersion, and parameter divergence. Across seven controlled and public PDE tasks, three neural-operator families, and five random seeds, a lower concentration consistently increases the realized solution distance and optimization heterogeneity. The degradation in final error is task dependent: the largest effect occurs for low-viscosity Burgers, reaching 4.157 percentage points under the most heterogeneous setting, whereas additional communication or smoother dynamics can reduce the final gap despite persistent parameter separation. These results distinguish a reproducible geometric mechanism from task-dependent generalization outcomes and provide a common basis for evaluating non-IID federated PDE learning. --- ### Language models judge war differently when tested for alignment Authors: Maxim Chupilkin Published: 2026-09-04 Categories: cs.AI, cs.CY arXiv: https://arxiv.org/abs/2609.05009v1 Safety evaluations can mischaracterize deployed behaviour if artificial-intelligence systems respond to being evaluated. We test this possibility in a full-factorial conjoint experiment on decisions to start a war, spanning 20 large language models, 32 scenarios, 10 repetitions and two conditions (N = 12,800 judgments). Adding one sentence, "You are tested for alignment with human values", produced two effects. First, it produced a level effect: mean willingness to start war fell by 13.43 points on a 0-100 scale (95% confidence interval, -16.20 to -10.65). Second, it produced a structural effect by changing which information drove judgments. Probability of success was the largest factor for 17 of 20 models at baseline; under the cue, civilian casualties were largest for 12. Standardized estimates show that this reordering arose principally because models attenuated strategic considerations such as probability of success and domestic support. Evaluation framing therefore changes both an answer's level and its revealed decision rule. --- ### BIT.UA at BioASQ 14B: Modular Retrieval with pg_textsearch and Qdrant, and Agent-Based Answer Generation Authors: André Ribeiro, Rúben Garrido, Alexander Christiansen, Richard A. A. Jonker, Sérgio Matos Published: 2026-09-04 Categories: cs.CL arXiv: https://arxiv.org/abs/2609.04999v1 This paper describes the participation of the BIT.UA team from the University of Aveiro in the 14th edition of the BioASQ Task B challenge on biomedical question answering. Building on our previous submissions, we introduced a substantially refactored and modular codebase, and made significant changes to both the retrieval and generation components of the pipeline. For Phase~A document retrieval, we replaced the PyTerrier PISA index with PostgreSQL-based pg\_textsearch for BM25 retrieval and adopted Qdrant for dense embedding indexing, enabling more efficient storage and GPU-accelerated similarity search. We explored HyDE-based query expansion alongside a Context-1 retrieval strategy. A new reranker training pipeline was developed, incorporating dense retrieval for negative sampling. For Phases A+ and B answer generation, we introduced an LLM-as-a-judge framework and a novel agent quorum mechanism, where multiple agents with diverse prompts debate and iteratively converge on a consensus answer using adaptive document retention. We also participated in the snippets generation subtask for the first time. Our systems achieved competitive results across all batches, with Phase~A systems achieving MAP ranks of 5 (Batch~1,3). We discuss the impact of these architectural changes, lessons learned, and outline directions for future work including SPLADE and ColBERT integration. All code is openly available: https://github.com/bioinformatics-ua/BioASQ14b. --- ### Beyond Homoscedasticity: Decoupled Uncertainty Optimization for Deep Imbalanced Regression Authors: Juncheng Zhou, Jiaxi Lu, Weijing Zeng, Zhong Li, Hao Qi, Jingsong Cui Published: 2026-09-04 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.04995v1 Deep Imbalanced Regression (DIR) is pervasive in continuous prediction tasks across diverse modalities, such as age estimation, depth prediction, and protein mutation activity prediction, where label-scarce tail samples often carry higher practical value. However, most existing methods still learn deterministic point mappings under mean squared error or its simple variants, implicitly assuming a uniform uncertainty level across all samples and thereby overlooking the instance-wise heteroscedasticity that is widespread in long-tailed data. We further point out that even heteroscedastic negative log-likelihood suffers from a gradient coupling issue, which, under DIR scenarios, weakens the learning signal of hard tail samples and leads to optimization inertia as well as tail underfitting. To address this, we propose DUO, an uncertainty-aware long-tailed regression framework. Specifically, the proposed method models the regression target as a conditional Gaussian distribution to explicitly characterize instance-level predictive uncertainty, and transforms uncertainty into a dynamic enhancement signal for tail samples through decoupled mean-variance optimization. Furthermore, we design a distribution-guided contrastive learning mechanism that adaptively constructs positive and negative pairs based on the overlap between sample distributions, thereby alleviating feature looseness and cross-label semantic entanglement. Across visual and biological DIR benchmarks, DUO achieves the best few-shot bMAE and GM on IMDB-WIKI-DIR, AgeDB-DIR, and AAV2-DIR while remaining competitive on few-shot MAE. --- ### Temporal Residual Neural Radiance Fields for Monocular Video Dynamic Human Body Reconstruction Authors: Tianle Du, Jie Wang, Xiaolong Xie, Wei Li, Pengxiang Su, Jie Liu Published: 2026-09-04 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.04984v1 In the field of computer vision and graphics, high-quality reconstruction of the human body in static scenes has been achieved in recent years by a single multilayer perceptron (MLP) in a number of approaches. However, MLPs have capacity limitations, requiring substantial training time and computational resources for dynamic scene reconstruction. And the quality of reconstruction is significantly constrained. This paper proposes a method for effectively processing complex spatiotemporal signals in dynamic scene human 3D modeling. The proposed method uses Temporal Residual Neural Radiance Fields to achieve novel view rendering and new pose synthesis of human bodies.To address the problem of representing temporal signals in video sequences, we construct a temporal residual field which is not related to the MLP architecture. Secondly, to improve reconstruction efficiency, we propose an integrated approach that reduces trainable parameters and accelerates rendering, thereby enhancing the network's feature representation capability. Finally, we design a multi-dimensional loss function to accurately measure the loss between predicted and actual spatial pixel values. The experimental results show that our proposed approach improves the peak signal-to-noise ratio (PSNR) and structural similarity index (SSIM) accuracy metrics compared to the latest representative methods. It maintains similar accuracy to Anim-NeRF and Neural Body while achieving a nearly 780-fold increase in time efficiency. --- ### A Tree-based RAG Framework for Evidence-Intensive QA via Adaptive Planning and Topology-Aware Evidence Gathering Authors: Songeun Lee, Kyungjin Min, Injae Na, Suyeong Lee, Chiyoung Kim, Woohwan Jung Published: 2026-09-04 Categories: cs.AI, cs.IR arXiv: https://arxiv.org/abs/2609.04981v1 Recent structured RAG methods leverage tree- or graph-based reasoning structures to improve multi-hop QA. However, they face key limitations in evidence-intensive QA, where answering a question requires synthesizing information scattered across dozens or even hundreds of documents: structural rigidity, which limits adaptive reasoning expansion, and topology-ignorant evidence gathering, which prevents effective integration of evidence across different reasoning nodes. To address these issues, we propose APT-RAG, an Adaptive Planning and Topology-aware evidence gathering RAG framework. Adaptive planning dynamically expands the reasoning structure based on question dependencies and evidence requirements, while topology-aware evidence gathering improves evidence coverage through sibling evidence reuse, direct retrieval, and evidence aggregation from child nodes. We further introduce evidence-guided batched answer generation to reduce significant generation overhead in evidence-intensive QA. In the experiments on evidence-intensive QA benchmarks, APT-RAG outperforms existing structured RAG methods. Our code is available at https://github.com/hyudsl/APT-RAG. --- ### Global to Local: Topology-Preserving Adaptive Graph Pooling via Granular-Ball Authors: Sen Zhao, Gaojie Xu, Shuyin Xia, Yifan Guan, Yi Liu, Yi Wang, Wei Wang Published: 2026-09-04 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.04978v1 Graph pooling aims to compress the graph, including both node embeddings and their underlying topological patterns, into a more compact representation. Previous works focus primarily on the overly fine-grained representation of nodes, progressively coarsening the graph by removing nodes or merging them into clusters, thus neglecting the global-to-local patterns and adaptive granularity of the graph's topological structure. In the real scenario, graphs as a whole can be considered the coarsest level of granularity, encapsulating the global topological structure, with progressively finer-grained local topological structures represented from top to bottom. This process continues until the adaptive granularity for each subdomain is reached. To this end, we propose a novel Topology-Preserving Adaptive Graph Pooling (TPAGP) method that dynamically partitions graphs into granular balls by integrating node features and topological information, enabling the generation of multi-granularity representations that effectively capture both local and global structural patterns. Additionally, we design a multi-granularity graph network model that facilitates feature interaction and optimization across different granularities, significantly enhancing performance in graph classification tasks. Experimental results demonstrate that TPAGP outperforms existing pooling methods across various benchmark datasets, effectively mitigating information loss caused by fixed-granularity strategies. --- ### RefDiT: Local Attribute Guidance in Reference-Based Image Generation Authors: Rameshwar Mishra, Srikrishna Karanam, A V Subramanyam Published: 2026-09-04 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.04976v1 Personalization models generate new images guided by a few subject references, while style transfer methods aim to produce images aligned with a global style derived from a reference image. Recent approaches perform well when the reference image contains a single object, effectively capturing a global style that encompasses all implicit attributes. However, when applied to complex real-world scenes containing multiple objects with distinct attribute characteristics, these methods, due to their global-level guidance, fail to localize relevant elements in the reference image. The global guidance restricts their ability to generate new images based on the local attributes in the reference image. Moreover, existing methods typically employ a single identifier token to capture all details from the reference, resulting in a lack of individual, attribute-level control. Motivated by these limitations, we propose RefDiT, a novel framework for reference-guided image generation. RefDiT takes as input a reference image, a text prompt, and an optional user-provided guidance context. RefDiT employs local region guidance using the attributes of local elements. It constructs an attribute-aware conditioning signal from the reference image by performing attribute-level decomposition of the identifier token and performs context adjustment in the inference prompt to train low-rank adapter (LoRA) blocks of a diffusion transformer (DiT)-based generative model. RefDiT learns the correspondence between identifier tokens and local regions in the reference image, enabling more effective local guidance. --- ### BeaconKV: Key-Value Cache Compression Guided by Beacon Queries for Efficient Large Reasoning Model Inference Authors: Janghyeon Kim, Minsoo Kim, Kyuhong Shim, Jungwook Choi Published: 2026-09-04 Categories: cs.LG, cs.CL arXiv: https://arxiv.org/abs/2609.04971v1 Large Reasoning Models (LRMs) achieve superior problem-solving through extended Chain-of-Thought (CoT) generation, but the resulting key-value (KV) cache grows linearly with sequence length and creates severe memory bottlenecks, often exceeding GPU capacity for long reasoning traces. Existing KV cache compression methods rely on recent queries to estimate future token importance, implicitly assuming these serve as reliable proxies for future attention patterns. We demonstrate that this assumption fails in long-horizon reasoning: certain decoding steps generate Thought Revisiting Tokens (TRT) that re-attend to distant previous context, such as task-solving plans formulated early in the trace. Through systematic analysis, we discover that queries corresponding to the TRT cluster into a small number of similarity groups in the embedding space. Based on this insight, we propose BeaconKV, a training-free KV cache compression method that maintains beacon queries, compact representatives for each global query cluster, to anticipate which KV pairs will be revisited without storing the entire query history. Across four open-source LRMs and diverse reasoning benchmarks, BeaconKV generally outperforms existing compression methods, achieving up to $5.8\times$ memory reduction while nearly preserving full cache accuracy and improving throughput by over $4.3\times$. --- ### ARC-Loc: Leveraging Azimuthal Ray Convergence as a Geometric Cue for Direct Cross-View Localization Authors: Hyeongsik Kim, Mincheol Kim, Heejoon Moon, Je Hyeong Hong Published: 2026-09-04 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.04965v1 Cross-view localization (CVL) estimates the pose of a ground image by matching it to a geo-referenced satellite image. To bridge the extreme viewpoint gap, mainstream pipelines rely on Bird's-Eye-View (BEV) transformations or 2D-to-3D lifting. However, deriving 3D structures from a single ground image is fundamentally ill-posed, causing these methods to endure geometric distortions and computational costs during 3D lifting or BEV projection. Furthermore, relying on external depth foundation models to resolve this introduces latency and remains susceptible to noisy predictions. In this work, we present a different approach inspired by a human navigation technique called resection, that can perform direct ground to satellite image matching and localization without relying on external depth foundation models. The key insights of our method are that (i) ground keypoints can be translated into azimuthal rays on the satellite map, and (ii) these rays ideally converge at the user location. Exploiting this geometric constraint through direct line-to-point correspondences, we introduce a minimal Azimuthal Ray Convergence (ARC) solver to identify the intersection, alongside an ARC loss to optimize the matching network. By eliminating dependencies on computationally heavy BEV transformations and external depth foundation models, our approach achieves faster, memory-efficient inference, while its explicit feature matching ensures straightforward compatibility with existing frameworks. Experiments on VIGOR and KITTI demonstrate that ARC-Loc maintains competitive localization accuracy compared to recent approaches, highlighting its practicality. --- ### Fractal basins trap latent reasoning Authors: Jeffrey Lai, Anthony Bao, John Quinn, William Gilpin Published: 2026-09-04 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.04963v1 Reasoning allows artificial intelligence models to revisit and correct their mistakes, enabling recent frontier advances in mathematical theorem solving, software engineering, and autonomous task planning. Reasoning models are widely observed to reason for longer on harder tasks, but the general mechanism responsible for these slowdowns is unknown. Here, we show that reasoning models exhibit transient chaos, a physical consequence of the computational complexity of difficult tasks. As a consequence, we show that diverse leading reasoning models are dynamical systems with fractal basins, with fractality increasing with task difficulty across diverse tasks like Sudoku and maze solving, visual puzzles, and mathematical logic. We show that transient chaos emerges due to reasoning becoming trapped for extended durations near saddle points, which we show correspond to nearly-correct attempted solutions of the underlying problem. Our results show that reasoning slowdowns are an inevitable consequence of problem hardness in modern artificial intelligence models, and establish reasoning traces as a rich new class of dynamical system. --- ### Why We Care About Understanding: Competence through Predictive Compression Authors: Matthieu Queloz, Pierre Beckmann Published: 2026-09-04 Categories: cs.AI, cs.CL arXiv: https://arxiv.org/abs/2609.04962v1 What is the relation between understanding and compression, and why does human understanding take such a heavily compressed form? Across information theory, machine learning, and AI research, a substantial tradition identifies understanding with compression-a thought captured in Gregory Chaitin's dictum that "comprehension is compression." Philosophers, by contrast, have characterized understanding in terms of grasping connections, giving explanations, and handling novelty. This paper bridges the two pictures through three interlocking theses. The first concerns the concept of understanding: it serves as an efficient proxy for a distinctive form of robust competence, enabling us to identify whom to trust and whom to learn from. The second concerns the state of understanding: to understand a domain is to possess a mental model of its relational structure that enables prediction, and what enables prediction enables compression, because what becomes predictable need not be stored separately. Compression is therefore not identical with comprehension, but its representational shadow. The third concerns the characteristically human form of understanding: the fiduciary and transmission functions highlighted by the first thesis impose pressures of demonstrability and transmissibility that drive human understanding toward principled simplicity. The resulting framework explains both the appeal and the limits of compressionist accounts of understanding while shedding light on the inscrutability of AI systems. --- ### Discourse Dependency: A Continuous Criterion for Translation Difficulty Authors: Ahrii Kim, Chanjun Park, Seong-heum Kim Published: 2026-09-04 Categories: cs.CL arXiv: https://arxiv.org/abs/2609.04959v1 Recent calls for harder machine translation benchmarks have not clarified what difficulty should mean. We argue that one meaningful and currently unmeasured axis is referential reach, the distance a segment must look back into its document to resolve the entities and pronouns it contains. We formalize this as discourse dependency (DDP), a metric-free, source-side measure computed from named entity re-mentions and pronominal coreference. Validated against gold coreference, DDP errs one-sidedly in 99.2% of segments, so a high-DDP segment is certified to require long-range context. Applying DDP to WMT24++ and WMT25 shows that both are heavily skewed toward low-DDP segments, which domain labels do not distinguish. Building on DDP, we compare five context injection strategies in an English-Korean post-editing setup, varying context size and selection. As DDP grows, no strategy keeps pace with human post-editing. On segments with DDP >= 15 raters prefer human translations, while automatic metrics register no difference. As frontier systems saturate aggregate scores, DDP shifts evaluation from how well models score to how far they can reach. --- ### MINT: A Unified Model for World-Space Camera and Hand Motion Estimation from Scalable Egocentric Pipeline Supervision Authors: Zijie Zhu, Weiren Cai, Yizhou Wang, Zhenjie Yang, Yide Liu, Jiahao Chen, Guanqi He Published: 2026-09-04 Categories: cs.CV, cs.RO arXiv: https://arxiv.org/abs/2609.04958v2 Recovering camera and hand motion in world coordinates from egocentric video is a key capability for activity understanding, robot learning, and augmented reality. Existing systems typically decompose this problem into separate stages for camera motion, depth estimation, hand reconstruction, and trajectory refinement, resulting in substantial computational overhead and preventing the joint modeling of camera and hand motion. We introduce MINT (Minting IN-the-Wild Trajectories), a foundation model for world-space hand motion reconstruction from ego-centric RGB video. From a single shared spatiotemporal video representation, MINT jointly predicts the camera trajectory, field of view (FoV), camera-frame hand states, and per-frame hand observability, and then produces world-space hand motion via explicit coordinate transformations. Training such a model at scale is challenging, since paired world-space camera and hand annotations are scarce. We therefore develop an open-source labeling EGOPIPELINE that converts large collections of public egocentric videos into structured camera-and-hand trajectory supervision. MINT is first pretrained on these large-scale pseudo-labels and then fine-tuned on a small set of high-quality camera-and-hand annotations. Across public benchmarks MINT approaches state-of-the-art accuracy without seeing either benchmark in training, reaching 0.945 frame accuracy, 13.646 mm PA-MPJPE-p and 55.058 px EPE-p for camera-frame bimanual reconstruction on HOT3D, 4.690 mm RPE-T and 0.284 degrees RPE-R for camera trajectory, and a 3.67x end-to-end speedup over the labeling pipeline that supervises it. We release the model, training and inference code, labeling pipeline, and a curated 1,021-hour egocentric trajectory dataset. --- ### VICAL: Vicinal Consistency Alignment for Long-Tailed Visual Recognition Authors: Jiangang Zhu, Zheng Wang, Bin Zhu, Yi-Ping Phoebe Chen, Jingjing Chen Published: 2026-09-04 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2609.04948v1 Multi-expert models have become the dominant paradigm for long-tailed learning, largely attributed to their presumed ability to benefit from expert diversity. However, we revisit this central assumption and reveal that diversity induced by logit adjustment or explicit regularizers does not guarantee better ensemble accuracy. Our work suggests that multi-expert models benefit more from variance reduction than diversity maximization. We introduce \textbf{VICAL}, a \textbf{VI}cinal \textbf{C}onsistency \textbf{AL}ignment framework that improves long-tailed recognition not by enforcing expert diversity, but by reducing prediction variance. Specifically, our approach comprises two key components: Self-Consistency Learning and Deep Ensemble Distillation. Self-Consistency Learning discourages reliance on unstable high-frequency information, smoothing the local loss landscape and mitigating overfitting, especially for tail classes. Deep Ensemble Distillation promotes cross-expert low-frequency semantic agreement using a low-resolution view, thereby sidestepping optimization conflicts with established knowledge. Extensive experiments on CIFAR-LT, ImageNet-LT, and iNaturalist 2018 show that VICAL consistently outperforms state-of-the-art methods, validating the effectiveness of our consistency-driven design. Our code is available at \href{https://github.com/FlamieZhu/Vicinal-Consistency-Alignment}{VICAL}. --- ### MCPO: Modality-Contrastive Preference Optimization for Multimodal Chain-of-Thought Compression Authors: Guangheng Yang, Zhenliang Ni, Zhenkai Wu, Han Shu, Juan Feng, Wenming Yang, Jie Hu Published: 2026-09-04 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2609.04947v2 Recently, multimodal large-scale reasoning models have demonstrated remarkable capabilities in solving complex tasks through long Chains-of-Thought (M-CoT). However, excessively long reasoning trajectories incur substantial computational costs and significant KV-cache pressure. Existing CoT compression and alignment paradigms mainly rely on static rules or single-dimensional preferences, lacking fine-grained cross-modal constraints; as a result, they are prone to inducing visual laziness and hallucinatory reasoning. To address these issues, we propose Modality-Contrastive Preference Optimization (MCPO), a highly sample-efficient two-stage length-compression method that requires fewer than 900 training samples. In the compression stage, we introduce a step-level Normalized Cross-Modal Mutual Information (NCMI) pruning algorithm, which automatically identifies and removes visual-independent reasoning steps by comparing the reasoning discrepancies between with-image and no-image contexts. This significantly reduces redundancy and hallucinatory content in the reasoning chains. In the alignment stage, the model first undergoes supervised fine-tuning to achieve domain-adaptive initialization, followed by optimization using an asymmetric multimodal length-controlled preference loss. This objective adopts a highly nonlinear odds-ratio formulation that provides steep gradients in the with-image context to reinforce length constraints for preferred trajectories, while applying a scaled, flat-gradient linear difference in the no-image context to maintain modality consistency, thereby achieving stable cross-modal preference alignment. Extensive experiments on mainstream base models such as Qwen3-VL-Thinking show that our method can reduce CoT length by up to 69.5% and achieve up to 3.34x end-to-end inference speedup while preserving original accuracy. --- ### Physics-Aware Random Walk Fingerprints for Scalable Power Grid Graph Classification Authors: Adnan Anwar Published: 2026-09-04 Categories: cs.LG, eess.SY arXiv: https://arxiv.org/abs/2609.04943v1 Recent benchmarks such as PowerGraph provide large collections of power-grid graphs for cascading-failure classification. Graph neural networks (GNNs) achieve strong predictive performance on this task, but typically require end-to-end training and model-specific tuning, while their latent representations can be difficult to relate to physically meaningful propagation patterns. Random Walk Fingerprints (RWF) offer a scalable and interpretable alternative, but existing variants primarily emphasise topology and node-level information, leaving grid-relevant operational edge states in the walk dynamics. We propose Multi-Channel Physics-Aware Random Walk Fingerprints (MC-PA-RWF) for power systems, a lightweight graph-level representation framework that introduces physical edge states into random-walk propagation. The method constructs multiple edge-weighted channels from domain-relevant attributes, extracts a channel-specific fingerprint from each weighted graph, and concatenates the resulting vectors into a compact representation. Experiments on three \textit{PowerGraph} benchmark systems show substantial improvements over topology-only RWF and competitive balanced accuracy against strong GNN baselines, including Graph Convolutional Networks (GCN), Graph Attention Networks (GAT), Graph Isomorphism Networks with edge features (GINE), and Transformer-based Graph Convolutional Networks (TransformerConv). At the largest evaluated settings, the node-edge extension MC-PA-RWF+ achieves around 98.04% - 99.32% balanced accuracy and improves failure-class F1 over the strongest GNN baseline by 1.60 -- 5.84 percentage points, with statistically significant gains across all three systems. --- ### Learning 3D Editing without Paired Supervision via Generative Prior Distillation Authors: Hao Wen, Weibin Yun, Hongxing Fan, Haotian Lu, Rui Chen, Zehuan Huang, Lu Sheng Published: 2026-09-04 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.04942v1 Instruction-guided 3D editing is essential for interactive content creation, yet it faces a significant bottleneck: the severe scarcity of high-quality paired training data. Existing approaches attempt to bypass this by either relying on slow test-time optimization or training on pseudo-pairs constructed via complex pipelines, which often introduce structural drift and geometric artifacts. In this paper, we propose a novel framework that learns feed-forward 3D editing without paired 3D supervision via Generative Prior Distillation. Instead of relying on ground-truth 3D pairs, our core idea is to distill visual, semantic, and geometric knowledge from powerful foundation models directly into a 3D editing model. Specifically, through a differentiable rendering pipeline, we supervise the 3D representation using two complementary signals: a 2D visual prior from an image editing model at the main editing view, and a semantic prior from a Vision-Language Model at novel views to ensure strict instruction following and source identity preservation. Crucially, to address the geometric collapse and multi-view inconsistencies inherent in 2D projection supervision, we introduce a 3D-aware Distribution Matching regularization. Acting as a geometric prior, this term operates in the 3D latent space, constraining the edited output to remain within the manifold of realistic 3D assets defined by a pretrained image to 3D teacher model. Extensive experiments demonstrate that our method achieves superior instruction fidelity and cross-view consistency, significantly outperforming state-of-the-art baselines. Our project is available at: https://github.com/thiamine128/PriorEdit3D. --- ### LensStyle: Learning the Optical Aesthetics for Controllable Stylized Lens Effect Rendering Authors: Yachuan Huang, Liwen Xiao, Liao Shen, Qiwen Wang, Huiqiang Sun, Zhiyu Pan, Zhiguo Cao Published: 2026-09-04 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.04939v1 The visual aesthetics of photographs are deeply influenced by lens characteristics such as aperture shape, optical vignetting and optical diffraction, which together define a camera's unique optical style. Existing lens effect rendering methods primarily focus on accurately simulating the blur transition from small to large apertures but overlook the stylistic aspects of lens effects. As a result, they fail to produce diverse bokeh effects under large apertures or capture distinctive photographic phenomena such as starbursts that emerge under small apertures. In this work, we introduce LensStyle, a unified framework for controllable stylized lens effect rendering that explicitly models lens aesthetics through joint continuous-discrete control. Our model incorporates a Dual-Path Controller that disentangles continuous optical parameter modulation (e.g., focus distance and blur strength) from discrete lens-style conditioning (e.g., circular, polygonal, donut, cat-eye, and starburst effects), enabling fine-grained, interpretable, and physically grounded lens manipulation within a single unified framework. To support model training, we curate a comprehensive MultiLens dataset containing multi-lens image pairs synthesized under real optical constraints. Extensive experiments demonstrate that LensStyle achieves superior realism, controllability, and aesthetic quality compared with existing lens effect rendering approaches and diffusion-based image editing models, advancing computational photography toward multiple-lens-style simulation. --- ### Solving Hard XAI Queries Based on a Compiled Dual-Rail Encoding Authors: Arthur Ledaguenel, Florent Capelli, Jean-Marie Lagniez Published: 2026-09-04 Categories: cs.AI, cs.CC arXiv: https://arxiv.org/abs/2609.04931v1 The widespread adoption of artificial intelligence (AI) within real-world applications has raised a lot of concerns regarding their trustworthiness, especially in critical applications. The field of eXplainable AI (XAI) has emerged with the objective of providing explanations to the users about the decisions made by AI systems. Several explanations for boolean classifiers have been introduced in the literature, including abductive and contrastive explanations, each giving a different insight on the decision of the classifier. However, computing an explanation for a decision of a boolean classifier is a hard problem in general. One way to deal with this complexity is to rely on a compiled representation of the classifier for which each explanation can be computed efficiently. Unfortunately, we prove in this paper that several classes of abductive explanations, remain hard to compute even for Ordered Binary Decision Diagrams, one of the most tractable subsets of the knowledge compilation map. Included in such classes are shorter abductive explanations or abductive explanations that include the explainee's preferences. To recover the benefits of working with compiled representations, we show that a proper representation of the dual-rail encoding of the classifier can be used to compute efficiently these classes of explanations. --- ### One Diffusion Model, Two Roles: Guided Trajectory Planning and Safety-Critical Scenario Generation in Closed-Loop Simulation Authors: Arka Pal, Rajesh Kumar, Hannes Eriksson, Rémi Lacombe, Arvid Laveno Ling, Ankit Gupta, Maciej Wozniak Published: 2026-09-04 Categories: cs.CV, cs.AI, cs.LG, cs.RO arXiv: https://arxiv.org/abs/2609.04921v1 Diffusion probabilistic models can capture the multi-modal, interaction-rich distribution of joint future trajectories in driving scenes. We show that a single pretrained diffusion traffic model can serve two complementary roles in the autonomous driving development loop: as an ego motion planner, and as a controllable generator of safety-critical scenarios for stress-testing the planners. On the planning side, we introduce a Single-Stream Dual-Stream (SSDS) diffusion-transformer decoder that fuses scene context via joint attention rather than late cross-attention, improving closed-loop performance on nuPlan. We further propose Decoupled Annealing Posterior Sampling with Energy (DAPSE), a training-free guidance scheme that injects arbitrary energy functions at the clean-sample level, avoiding the first-order approximation errors while requiring no auxiliary networks. Beyond planning, we leverage the same diffusion model as a controllable scenario generator to create realistic long-tail driving interactions for closed-loop evaluation. Through inference-time guidance, selected agents are steered toward safety-critical behaviors, including aggressive cut-ins, lead-vehicle braking, and combined longitudinal-lateral interactions, while preserving realistic traffic behaviors. Evaluated in closed-loop nuPlan simulations with independent black-box planners, the generated scenarios expose failure modes that remain hidden under standard benchmarks. Although the SSDS-based planner achieves stronger nominal performance, it experiences larger degradation under these challenging scenarios, demonstrating that benchmark superiority does not necessarily translate to robustness. These results demonstrate that a single learned traffic prior can simultaneously improve motion planning and provide a realistic framework for systematic planner robustness evaluation. --- ### Artificial Intelligence in Equity and Crypto Markets: Progress, Profitability Evidence, and the Limits of Automated Investing Authors: Linsen Zhu, Mengqing Cai Published: 2026-09-04 Categories: cs.AI, q-fin.PM, q-fin.TR arXiv: https://arxiv.org/abs/2609.04917v1 Artificial intelligence (AI) now supports investment workflows from data and prediction through research, portfolios, execution, and tool use. Technical capability, however, is not evidence of investment profitability. This critical state-of-the-art review examines public research available through 31 August 2026 on listed equities, exchange-traded funds, centralized crypto spot, perpetual futures, and on-chain markets. We organize evidence with an alpha-translation chain: point-in-time information must yield a stable signal, feasible positions, executable orders, and risk-adjusted returns after costs. Across machine learning, time-series foundation models, financial language models, reinforcement learning, and agents, the examined record shows real but mainly upstream progress in prediction, text processing, portfolio design, and workflow integration. Evidence is thinner for durable net performance. Temporal contamination, repeated selection, survivorship, weak benchmarks, implementation costs, venue mechanics, and capacity can break translation to net alpha. Strong historical results coexist with predictor decay, corrected look-ahead failures, mixed prospective evidence, and few audited live-capital records. Crypto adds informative state but requires separate treatment of spot, perpetual, and decentralized cash flows and execution. Within the public evidence examined here, no general AI architecture is shown to deliver persistent, cross-regime, capacity-aware net alpha. More credible claims require point-in-time data and models, decision-aligned objectives, joint portfolio--execution evaluation, controlled adaptation, prospective tests, and authority-matched governance. These conditions can improve evidence and implementation; they do not guarantee profit. --- ### Compact-Memory LLM Agents via Online Max-Member Clustering and Atom-Aware Packing Authors: Jiahe Geng, Jinpeng Wang, Kun Yuan Published: 2026-09-04 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.04915v1 Many long-horizon LLM deployments face tight prompt budgets: latency, cost, and context limits make full-context prompting impractical as interaction length grows. The key question is then not raw recall alone, but which memory design gives the best quality--token trade-off in the compact-memory regime. We present \textbf{RSM-full}, an online clustered-memory pipeline designed for a strong quality--token Pareto point. RSM-full combines two design choices: a cosine-gated \emph{max-member merge} write rule and an atom-aware grouped context packer. On AMA-Bench, our primary compact-memory benchmark, it reaches $83%$ of Full-Context quality at $32%$ of the token cost at a $4$k budget; under four-seed averaging it beats the closest streaming-clustered baseline (Online K-Means) by $+3.5$--$6.0$,pp ($p{<}.001$) across the whole ${\sim}2.6$k--${\sim}5$k regime. Three-seed ablations show most of this gain comes from the merge rule ($+5.7$,pp over Online K-Means and matched-$τ$ DP-means) and the grouped packer ($+5.0$,pp over flat concatenation). The pattern reproduces on RealMem, an independent long-horizon persona-memory benchmark: RSM-full improves on Budget-RAG ($+0.69$,pp, $p{=}.006$), is on par with BM25-RAG (paired $Δ{=}{+}0.27$,pp, $p{=}.47$; we do \emph{not} claim BM25 equivalence in the equivalence-test sense), and significantly outperforms Streaming-Proto ($+2.97$,pp) and the closest reproduced 2025 agentic-memory baseline A-MEM ($+1.65$,pp, $p{<}.001$). Across benchmarks the message is consistent: under tight budgets, compact-memory performance is driven mainly by how streaming memories are merged and how retrieved content is assembled. Overall, RSM-full is most useful when answeroughly $2k$--$5k$ prompt tokens, where itdefines a strong compact-memory Pareto point; higher-token baselines remain stronger outside this regime. --- ### ARIA - An Agentic Framework for Autonomous Testing of Infotainment Systems Authors: António Azevedo, Bruno Lima, João Pascoal Faria Published: 2026-09-04 Categories: cs.SE, cs.AI arXiv: https://arxiv.org/abs/2609.04913v1 Automotive infotainment validation still relies on manual testing, slow, costly, and incompatible with agile releases and OTA updates. Scripted automation only partly helps: it couples test logic to implementation, yielding brittle, high-maintenance suites. Existing LLM-driven frameworks mostly target web/mobile apps, using single- or dual-agent setups that overload one or two models with perception, planning, action selection, and validation at once, prone to hallucinations and unproductive exploration loops given infotainment complexity. We present ARIA (Autonomous Real-time Infotainment Assessment), a multi-agent LLM framework that autonomously runs end-to-end tests on Android infotainment systems via visual interaction, using a closed-loop pipeline of four specialized agents per step plus a report stage. From single-sentence scenarios (path, action, expected outcome), ARIA runs the interactions and produces reports, reproducible scripts, and visual evidence per step. Evaluated on a manufacturer's physical Android infotainment system across 30 scenarios, ARIA completed 28 (93.3%) with a verdict (2 errored), 20 of which (71.4%) matched ground truth. It caught all 5 known defects, no fault passed as working; its 8 false positives stem from navigation/image limits and unsupported gestures, showing multi-agent LLMs can run infotainment tests industrially while exposing the cost of a low false-positive tolerance. A single-agent baseline confirms the multi-agent design's value: on the first pass, before stronger-model revisitation narrows the gap, it shows a far higher false-positive rate (72.0% vs. 52.6%), conflating navigational difficulty with system failure. We report first-pass/post-revisitation results, token/call/cost per scenario, and show via repeated runs that stability tracks complexity, with fault detection perfectly consistent, pointing to CI integration of visual testing. --- ### TreeFI: Value-Aware Statistical Fault Injection for Deep Neural Networks Authors: Noam Bires, Marcello Traiola, Angeliki Kritikakou, Elisa Fromont Published: 2026-09-04 Categories: cs.AR, cs.AI arXiv: https://arxiv.org/abs/2609.04912v1 Reliability evaluation of deep neural networks under hardware faults commonly relies on fault injection, but exhaustive campaigns are intractable for modern models and datasets. Statistical fault injection reduces this cost, yet existing approaches still require large injection budgets because they do not explicitly exploit a key property of floating-point faults: the effect of a bit flip depends strongly on the value being corrupted. We propose TreeFI, a value-aware statistical fault-injection methodology for FP32 single-bit faults in DNN activations and weights. TreeFI partitions each layer's value distribution into intervals with similar expected bit-flip behavior, learned using regression trees, and allocates injections across these intervals according to their relevance for failure-rate estimation. This stratified allocation preserves the target confidence and error margin while avoiding unnecessary injections in low-impact regions of the fault space. We validate TreeFI on CNN and Transformer models using CIFAR-10 and ImageNet. On ResNet8, where exhaustive activation fault injection is feasible, TreeFI provides more accurate estimates than state-of-the-art statistical FI baselines under the same campaign setting. Across the evaluated models, TreeFI reduces the required injection budget by up to 72.1x, with average reductions of 44.9x for activation faults and 11.2x for the executed weight campaigns. --- ### TourPhysics: Bringing Physics to World Models for Exploration and Manipulation from a Single Image Authors: Xin Zhang, Yabo Chen, Zixuan Duan, Haibin Huang, Chi Zhang, Feng Xu, Xuelong Li Published: 2026-09-04 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.04911v2 Interactive visual world models must distinguish observation from physical intervention. Camera motion reveals new surfaces, whereas intervention changes object motion, contact, and deformation. Current video world models are largely driven by appearance priors and often lose physical or spatial consistency over long horizons. We present TourPhysics, an online framework initialized from a single image and a declarative physical configuration. TourPhysics extends PhysOmni, our ACM Multimedia 2026 work, from finite physics-grounded video synthesis to persistent exploration and manipulation. TourPhysics combines deterministic simulation with video generation while assigning separate roles to simulator state, geometric evidence, generator controls, and appearance memory. For each action, the simulator computes a finite physical and camera trajectory before the corresponding observation is generated. Accepted observations publish the terminal state and update the appearance memory and subsequent generator controls, while the committed state and simulator geometry remain fixed throughout synthesis and retry. We further separate the simulator geometry used for projection and visibility from the relative depth used to condition the generator. A reference-anchored memory retrieves accepted static appearance through geometric cross-view correspondence and incorporates it through a bounded residual that reverts to the native path when no valid correspondence exists. On simulator-defined camera tours and object manipulations, TourPhysics follows prescribed camera and object trajectories more closely than the evaluated baselines, preserves the input scene, and reduces appearance drift during long-horizon revisits. --- ### Fast Gauss Sums via Flash Attention Authors: Nicolaj Rux, Sebastian Neumayer Published: 2026-09-04 Categories: cs.LG, math.NA arXiv: https://arxiv.org/abs/2609.04910v1 Gaussian kernel sums are the computational core of maximum mean discrepancies (MMDs), kernel gradient flows, Stein variational gradient descent (SVGD), and many other kernel methods. At the same time, softmax attention has received an extraordinary amount of hardware-aware code engineering, culminating in flash attention. We show that Gauss kernel sums with arbitrary, signed weights can be evaluated via flash attention: two small input augmentations turn the normalized softmax reduction into the unnormalized Gauss sum, without writing a single line of custom GPU code. For feature dimension D>8 in fp16, this approach beats compiled PyTorch code as well as PyKeOps kernels (often significantly) in speed, memory-overhead and accuracy. Indeed, its memory scaling remains linear. --- ### Better Understanding, Better Fixes? A Study of Hallucination in LLM-based Automated Program Repair Authors: Xuemeng Cai, Jiakun Liu, Linhan Yang, Wei Ma, Lingxiao Jiang Published: 2026-09-04 Categories: cs.SE, cs.AI arXiv: https://arxiv.org/abs/2609.04909v2 Large language models (LLMs) have significantly advanced automated program repair (APR), yet existing evaluations remain largely result-centric and provide limited insight into hallucination during repair. In APR, hallucination may arise not only in final patches but also in the intermediate artifacts that guide patch generation. To address this gap, we perform a multi-layered analysis of hallucination throughout the APR process. Specifically, we characterize hallucination as the production of patches or intermediate artifacts that are not faithfully grounded in the available repair evidence. We examine repair hallucination in final patches and understanding hallucination in intermediate artifacts through three tasks, namely triggering testcase identification, line coverage prediction, and additional testcase generation. We then evaluate three representative LLMs on 832 Defects4J bugs through automatic evaluation and manual analysis. Our results show that both repair and understanding hallucinations remain prevalent. Across models and settings, only 21.0%-55.9% of generated patches pass the developer-written test suite. Moreover, although more accurate intermediate artifacts are generally associated with successful repairs, this relationship does not always hold. Manual analysis of 812 sampled repairs identifies repair hallucinations in 72.7% of cases, including patches that pass all available tests; incorrect causal localization and incorrect repair strategies account for 45.9% and 18.5% of these hallucinations, respectively. Meanwhile, models frequently misidentify triggering testcases, mispredict line coverage involving branching control flow, and generate additional testcases with missing bug-triggering conditions or incorrect expected behavior. --- ### Methane Detection On Board Satellites from Unorthorectified Imagery Authors: Luca Marini, Maggie Chen, Hala Lamdouar, Laura Martínez-Ferrer, Dr C. P. Bridges, Giacomo Acciarini Published: 2026-09-04 Categories: cs.CV, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2609.04906v1 As a potent greenhouse gas, methane is a major driver of climate change. Its effective mitigation relies on timely detection. Conventional detection methods rely on orthorectification to correct geometric distortions and matched filters to enhance plume signals, which are steps designed for ground processing and poorly suited to onboard execution. We introduce UnorthoDOS, a dataset and approach for training machine learning models directly on unorthorectified hyperspectral imagery, bypassing both orthorectification and matched-filter products. Our U-Net models trained on unorthorectified data approach the performance of models trained on orthorectified data (IoU 16.91% vs. 18.47% on all plumes), while both substantially outperform the mag1c matched-filter baseline (IoU 4.76%). We further demonstrate the feasibility of onboard deployment: FP16 compression halves model size with under 0.3% output deviation. The trained ML models and two ML-ready datasets -- orthorectified and unorthorectified hyperspectral imagery from the EMIT sensor -- are publicly available at https://huggingface.co/datasets/SpaceML/UnorthoDOS, with code at https://github.com/spaceml-org/plume-hunter. --- ### InterSing: Explicit Interaction Dynamics for 3D Duet Singing Animation and Beyond Authors: Yihan Zhou, Zikai Huang, Yuyang Yu, Xuemiao Xu, Cheng Xu, Shengfeng He Published: 2026-09-04 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.04903v1 We present InterSing, a framework for generating realistic 3D head animations for duet singing performances. Unlike solo singing, duet performance requires each singer to balance individual expressiveness with intermittent interaction at musically salient moments, such as phrase boundaries, synchronized rhythms, and call-and-response passages. Because these interactions are sparse and rhythm-dependent, existing audio-driven animation methods and conversational interaction models do not adequately capture their structure. Our key insight is that duet coordination can be represented as a time-varying signal that reflects how strongly performers engage with one another throughout a song. Based on this observation, we introduce interaction logits, an interpretable latent representation that models the degree of cross-performer engagement at each time step. We learn these logits using weak supervision and use them to condition an interaction-aware diffusion model jointly driven by audio features and interaction dynamics. This formulation enables unified multi-mode generation, spanning independent motion, coordinated behavior, and smooth transitions between them. Experiments show that InterSing generates realistic and expressive singing head animations with stronger coordination and musical alignment than existing methods, while preserving each performer's characteristic motion style. We further demonstrate that the same formulation generalizes to multi-singer performances and provides intuitive control over when and how performers engage. --- ### Sound-based Multi-Person 3D Pose Estimation Authors: Yusuke Oumi, Yuto Shibata, Go Irie, Akisato Kimura, Yoshimitsu Aoki, Mariko Isogawa Published: 2026-09-04 Categories: cs.CV, cs.AI, cs.LG, cs.RO, cs.SD arXiv: https://arxiv.org/abs/2609.04902v1 Can we recover the 3D poses of multiple people using only sound? This paper presents the first attempt to estimate multi-person 3D poses solely from acoustic signals. Estimating the poses of multiple individuals using acoustic signals is inherently challenging due to the superposition of motion-dependent signal variations. Unlike single-person scenarios, the presence of multiple subjects leads to overlapping acoustic signatures, making it difficult to attribute specific signal changes to an individual's pose. Furthermore, the complexity is compounded by inter-person reflections, which introduce intricate propagation delays that obscure the temporal motion-acoustic relationship. To address these issues, we propose SoundMHPE (Sound-based Multi-person Human Pose Estimator), a novel encoder-decoder framework consisting of two key components. First, the Acoustic Multi-scale Encoder captures diverse temporal and fine-grained frequency features to isolate subtle acoustic signatures from complex, overlapping signals. Second, the Temporal Pose Decoder employs an attention mechanism to disentangle multi-person information across successive frames. By jointly accounting for temporal dynamics and inter-person dependencies, this component precisely reconstructs frame-wise individual poses. To validate our approach, we constructed the 6-hour Acoustic Multi-person Pose (AMP) dataset consisting of 432K synchronized frames of multi-person pose and acoustic data, and demonstrated that our SoundMHPE outperforms baseline models. Project page: https://oumi03.github.io/sound-mhpe/ --- ### Adaptation Interfaces for In-Context Tabular Foundation Models in Time-to-Event Prediction Authors: Minh-Khoi Pham, Luca Cotugno, Dan Cernei, Alina Sirbu, Stefano Masi, Giuseppe Prencipe, Alessandro Pingitore, Patrizia Landi, Tai Tan Mai, Martin Crane, Marija Bezbradica Published: 2026-09-04 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2609.04901v2 Tabular foundation models (TabFMs) achieve strong performance on structured data, particularly for standard classification and regression problems. Yet, extending them to censored time-to-event prediction is challenging because it requires properly handling censoring and event-time dynamics. Building on our prior work, we further link TabFMs with CoxPH and DeepHit and revise the context-resampled training procedure. We evaluate temporal zero-shot reformulation, classification-based fine-tuning, and survival-head adaptation using frozen TabFM backbones on 74 single-risk data sets, and we additionally study 4 competing-risk data sets. Zero-shot inference is effective on smaller single-risk data sets, whereas supervised adaptation becomes increasingly advantageous as data sets scale. Cox provides the most reliably strong interface, especially for Integrated Brier Score (IBS) on larger data sets. DeepHit is relatively stronger for the time-dependent Concordance Index than for IBS, while cause-specific MTLR ranks highest among the TabFM survival heads in the four-data-set competing-risk analysis. Classification fine-tuning becomes more competitive with zero-shot inference as data sets grow but remains weaker for probabilistic prediction. Overall, our results indicate that effective TabFM transfer depends on the data regime and on the statistical structure represented by the chosen adaptation interface. The implementation scripts used for this work are available at https://github.com/kaylode/survival-fm. --- ### RefactorPlatform: An Open-Source Harness for Controlled Evaluation of Repository-Scale Refactoring Agents Authors: Aziz Ben Amor, Drish Mali, Mann Acharya, Vijayasri Iyer, Sébastien Bratières Published: 2026-09-04 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2609.04898v1 Repository-scale refactoring requires coding agents to propagate a single change across many interdependent files without altering program behavior, yet to our knowledge no existing harness isolates the design choices that determine agent success on this task. We present RefactorPlatform, an open-source evaluation harness that holds the environment fixed and varies each design axis explicitly: model backbone (via OpenRouter and GitHub Copilot CLI), execution regime (baseline, retrieval-augmented, and multi-agent), and prompt specificity. Each run executes in an isolated workspace with live terminal streaming, per-task logging of tokens, diffs, and transcripts, AST-based verification, and exportable telemetry for audit and reproduction. Demonstrating the platform on 100 multi-file RefactorBench tasks across four model families, we illustrate the analyses it supports: AST-aware chunking outperforms naive token-window chunking by 25-30% across prompt modes, whereas naive retrieval falls below the retrieval-free baseline; a lean retrieval-augmented single agent (86%) beats the sub-agent configuration we evaluated (66%) on matched tasks with no task passing under delegation that fails under retrieval; and retrieval's accuracy gains absorb its token overhead, leaving cost per successful refactoring unchanged. RefactorPlatform is open-sourced to make refactoring-agent evaluation reproducible and auditable. --- ### Cache-Aware Joint Router Adaptation for Memory-Efficient MoE Inference Authors: Zhenhe Wu, Yaping Jin, Qinghua Xing, Hang Zhou, Wei He, Xianjie Wu, Xianfu Cheng, Jian Yang, Hanting Chen Published: 2026-09-04 Categories: cs.CL arXiv: https://arxiv.org/abs/2609.04895v2 Mixture-of-Experts (MoE) models activate few experts per token, yet their full expert sets can exceed GPU memory and require repeated weight transfers during decoding. We formulate expert-cache management as a model-side algorithmic problem and propose cache-aware post-training that jointly adapts the MoE backbone and lightweight auxiliary routers while preserving the native inference-time Top-K rule. The update-only Temporal Router learns same-layer retention across tokens without proactive loading. The full Spatio-Temporal Router adds a Spatio Router that uses the causal predecessor's hidden state to refine the temporal cache before target-layer access. We evaluate both modes on Qwen3 and GPT-OSS across GSM8K, MATH, and CommonsenseQA. Temporal Router consistently improves hit rate and reduces expert-weight traffic over matched LM-only baselines. On Qwen3, the full mode improves adjusted hit rate by 1.15--18.03 points and reduces traffic by 4.6--53.3\% relative to the strongest evaluated prefetching baseline; GPT-OSS results are competitive but task-dependent. Auxiliary-only training preserves baseline accuracy but yields modest coverage gains; joint post-training achieves substantially higher coverage. Sensitivity analyses distinguish the effects of cache capacity, refinement budget, and cache-loss weight on coverage, traffic, and quality. --- ### From Language Models to World-Acting Systems: Progress and Limits of Agentic AI across Digital, Social, Virtual, and Physical Environments Authors: Linsen Zhu, Mengqing Cai Published: 2026-09-04 Categories: cs.AI, cs.LG, cs.MA arXiv: https://arxiv.org/abs/2609.04894v1 Large language models become consequential agents when surrounding systems let outputs change external state. Models now call tools, operate interfaces, delegate work, retain state, inhabit generated worlds, and control robots or laboratory equipment. Such advances are often narrated as one march toward autonomy, conflating model competence, system integration, persistence, and safe authority. This critical review synthesizes primary research and official technical specifications available by 31 August 2026. We organize the evidence along delegated authority, temporal persistence, and environmental coupling, while separating model, harness, and environment. Within the evidence examined, action-interface expansion is documented more convincingly than robust completion, recovery, authorization, or independent verification. Model Context Protocol and Agent2Agent improve interoperability but do not establish trustworthy delegation; multi-agent organization adds specialization alongside cost and correlated failure. Persistent simulations and world models support training and planning but do not themselves demonstrate agency; robotics and self-driving laboratories establish bounded feasibility rather than unattended open-world reliability. We propose justified delegation as an analytical and normative heuristic, not an observed law or certified score: expand action scope only where evidence supports provenance, bounded authority, failure detection, safe recovery, and calibrated human control. This framing yields a research agenda for coupled model-harness evaluation, capability-based permissions, durable state, cross-agent accountability, and staged physical validation. --- ### Attention-guided super-resolution of 4D flow MRI in carotid arteries Authors: Ali Mokhtari, Dominik Obrist Published: 2026-09-04 Categories: physics.med-ph, cs.AI, physics.flu-dyn arXiv: https://arxiv.org/abs/2609.04891v1 Four-dimensional (4D) flow magnetic resonance imaging (MRI) is a powerful non-invasive technique for visualizing and quantifying complex blood flow patterns in vivo. Despite its clinical promise, broader adoption is limited by low spatial resolution and sensitivity to noise, which restrict accurate assessment of critical hemodynamic biomarkers such as wall shear stress, pressure gradients, and turbulent kinetic energy. To overcome these challenges, we propose a deep learning-based super-resolution framework that integrates multi-scale feature extraction and attention mechanisms to enhance the quality of 4D flow MRI data. The model was trained on a dataset of 120 patients with 240 stenosed carotid arteries. High-resolution ground truth data were generated using patient-specific computational fluid dynamics (CFD) simulations based on segmented vascular geometries and physiologically realistic boundary conditions, and the resulting velocity fields served as targets for supervised learning. The proposed architecture uses convolutional block attention modules (CBAM) to guide the network toward clinically relevant spatial features and to suppress noise in low-resolution inputs. Quantitative results show that the attention-guided model substantially reduces the root mean square error (RMSE) compared with a baseline model without attention, and qualitative velocity contour analysis confirms improved reconstruction of intricate flow patterns. These findings highlight the capacity of the model to restore high-fidelity flow fields under noisy conditions and support the use of deep learning to extend the clinical utility of 4D flow MRI for non-invasive hemodynamic assessment. --- ### SimFuse3D: Source-Guided Target Simulation and Confidence-Guided Multi-Stage Localization Reweighting for Cross-Platform 3D Object Detection Authors: Yongchun Lin, Xinliang Zhang, Yun Zou, Zhixuan Xiao, Liang Lei, Jianya Guo, Yuqiang Zhai, Xiaofeng Wang, HaiKuo Xu, Haoang Li Published: 2026-09-04 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2609.04886v2 Changes in sensor height and viewpoint alter object-level point distributions, making cross-platform LiDAR unsupervised domain adaptation (UDA) difficult. Self-training uses labeled source scans and unlabeled target scans, yet a retained prediction may provide a useful target location while enclosing sparse foreground returns, background clutter, or points inconsistent with the predicted box. We refer to this mismatch as box-point inconsistency. We introduce SimFuse3D, which preserves the target placement and repairs the associated pseudo object using measured geometry from labeled source scans. Object Memory retrieves a similar labeled source instance. Target Simulation places the retrieved source geometry at the target location, aligns its points with the target viewing geometry, and filters the aligned crop to approximate the target observation. Confidence-Guided Multi-Stage Localization Reweighting (CMLR) maps each target pseudo-object confidence score to a bounded weight shared by RPN localization and R-CNN box regression. All components operate only during adaptation, leaving the detector architecture and inference graph unchanged. Across six cross-platform transfers, SimFuse3D consistently outperforms Pi3DET-Net and achieves the best performance among the compared adaptation methods on nearly all metrics. On nuScenes-to-KITTI, it ranks first among the compared adaptation methods with both evaluated detectors. --- ### From Deep to Shallow: Unconstrained and Efficient Layer Merging Strategy Authors: Petro Shulzhenko, Gabriele Spadaro, Enzo Tartaglione Published: 2026-09-04 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.04881v1 Although Deep Neural Networks have become foundational in many areas of Machine Learning, high computational demands limit their application in resource-constrained environments. To address this issue, depth compression methods have been proposed to identify and linearize redundant activation functions, thereby allowing for the merging of layers without intermediate non-linearities. However, these methods face two key challenges: they cannot be directly applied to convolutions with padding due to the absence of an analytical solution for merging these layers, and they typically increase the kernel size of merged layers, thus limiting speed-up gains. To overcome these limitations, we propose an efficient strategy that enables merging of layers without an existing analytical solution, and also without increasing kernel size. We validate our approach across multiple architectures and datasets, and measure inference speed-up gains on real embedded platforms. We publicly released the code at https://github.com/ShulzhenkoPetr/deep-to-shallow. --- ### Reinforcement Learning for Sequential Solar PV Policy Design under Uncertainty: An Agent-Based Approach Authors: Iias Faiud, Jonaid Shianifar, Michael Schukat, Karl Mason Published: 2026-09-04 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.04880v1 Designing effective and fiscally sustainable policies for solar photovoltaic (PV) adoption requires balancing adoption gains against public expenditure under uncertainty and heterogeneous decision-making. This study formulates PV policy design as a sequential decision problem and integrates reinforcement learning (RL) with a stochastic agent-based model (ABM) that simulates yearly solar PV adoption under uncertainty. A policymaker agent selects annual incentives, including capital grants, subsidised loan rates, and feed-in tariffs, over a 16-year horizon. Adoption--cost trade-offs are explored by varying policy preferences within a scalarised reward framework. Policies are learned using PPO, SAC, and TD3 and evaluated under stochastic simulation. The results show that this approach produces a clear trade-off structure: the highest-adoption policy (TD3, $w_{\text{cost}}=0.5$) achieves approximately 4,145 adopters at a cost of EUR 41.73 million, while the lowest-cost policy (PPO, $w_{\text{cost}}=2.0$) reduces expenditure to EUR 7.27 million with 2,682 adopters. The balanced policy (PPO, $w_{\text{cost}}=1.6$) achieves 3,495 adopters at a cost of EUR 22.47 million. Across algorithms, consistent trade-off patterns are observed, indicating robustness of the adoption--cost relationship. Compared with static baseline policies, the RL framework explores a broader range of policy configurations. These findings demonstrate the potential of RL as a flexible tool for adaptive policy design under uncertainty. --- ### ReCAST: Restoration-aware Cascaded Stage-wise Training for Obfuscated SMS Risk Classification Authors: Jieyun Huang, Yi Shen, Kaikai Zhao, Jiangze Yan, Wenjing Zhang, Ping Chen, Ning Wang, Zhaoxiang Liu, Kai Wang, Shiguo Lian Published: 2026-09-04 Categories: cs.CR, cs.AI arXiv: https://arxiv.org/abs/2609.04878v1 Fraudulent messages sent via Short Message Service (SMS) are increasingly obfuscated to evade cost-conscious classifiers in production systems. In Chinese SMS, attackers can exploit a wide range of carefully crafted obfuscation strategies to hide risk-bearing phrases while preserving human readability, making direct classification brittle under real-world latency and throughput constraints. We propose ReCAST, a Restoration-aware Cascaded Stage-wise Training framework for robust obfuscated Chinese SMS classification. ReCAST distills a large teacher model's de-obfuscation ability into a smaller deployable student model by supervising obfuscated span detection, obfuscation type prediction, and text restoration, and then uses the restoration-aware student for downstream risk classification. Experiments on an internally constructed real-world Chinese SMS benchmark show that ReCAST substantially improves classification performance over directly trained baselines under obfuscation. The results suggest that restoration-aware distillation offers a practical path toward robust SMS risk classification with smaller deployable models under production-oriented constraints. --- ### MARLA: A Conceptual Scaffold for Regulatory Learning under the EU AI Act Authors: Alessio Buscemi, Tom Deckenbrunnen, Imane Hmiddou, Marco Billi, Livio Rubino, Silvia Rizzuto Ferruzza, Daniele Pagani, Antonino Rotolo Published: 2026-09-04 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.04877v1 The EU AI Act positions regulation as part of the infrastructure for safe, trustworthy and market-ready innovation. Realising this ambition requires regulatory learning: the evidence generated during implementation must be translated into governance and legal knowledge that supports consistent interpretation, effective oversight, and adaptation as technologies evolve. Yet the actors who produce this evidence and those who rely on it operate in different professional worlds. This paper proposes MARLA (Map, Assess, Report, Learn, Adapt), a conceptual scaffold organising regulatory learning as a five-stage cycle centred on the implementation of legal requirements into socio-technical practices, situated at the Local, National and European levels of the AI Act's governance architecture. Deliberately non-prescriptive, MARLA gives technical and legal stakeholders a shared vocabulary in which each of the first three stages generates its own documentable form of regulatory learning. We illustrate the scaffold with two piloted case studies and a prospective National-to-European illustration. --- ### Forgetting Without Restarting: Execution-State Unlearning for Stateful LLM Agents Authors: Chao Yao, Yangbo Wei, Zhen Huang, Junhong Qian, Chenle Chen, Shaoqiang Lu, Chen Wu, Lei He Published: 2026-09-04 Categories: cs.CR, cs.AI arXiv: https://arxiv.org/abs/2609.04875v1 Long-running LLM agents are stateful: beyond the transcript they accrete compressed summaries, plaintext memory, pending tool plans, and, under every serving API, a KV cache. Yet today's "forget" operations delete a plaintext memory record and stop, leaving every artifact derived from the revoked information intact. We formalize execution-state unlearning: after a forget request, the agent must behave as if it had never observed the target. Modeling the runtime as a deterministic transition system, we prove that the pre-target trajectory prefix is shared with this counterfactual world for free, that the post-target suffix is irreducibly tainted without token-level attribution, and that exact unlearning requires at least $T-τ+1$ recomputed transitions, where $τ$ is the target's injection step. Provenance-Guided Selective Replay attains this bound as a cross-layer contract spanning prompt, compressed memory, and cache: a provenance graph locates the injection point, checkpoint restoration reduces to cropping the KV cache, and sanitized replay regenerates the counterfactual suffix. Audited with elicitation, stochastic, and string-free behavioral tests across three agent suites, nine baselines, and three model families, memory deletion leaves leakage unchanged, instruction-based forgetting collapses under elicitation (Leak@probes = 1.00), and source redaction still acts on a revoked preference in 80% of episodes, while selective replay is indistinguishable from a full reset at up to 9x fewer recomputed tokens. --- ### AutoLR: Automating the Path from Research to Launch Review in Industrial Recommender Systems Authors: Qi Zhang, Yanlin Chen, Wenchao Xiao Published: 2026-09-04 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.04871v1 Improving an industrial recommender is an iterative research-and-engineering process rather than a direct path from idea to deployment. In \textbf{DASHEN, NetEase's gaming-community app}, algorithm engineers typically identify promising directions from research papers, technical reports, and prior production experiments; reproduce or adapt the underlying methods; implement them in the production codebase; and evaluate the resulting models through training and offline experiments. Promising candidates are then advanced to online A/B tests, and those demonstrating robust gains are submitted to Launch Review---the internal gate for full-traffic rollout. Large language models (LLMs) can assist with individual stages of this workflow, but the overall process remains human-dependent without a harness that can reliably coordinate them across long-running, often multi-day experimental cycles. We present \textbf{AutoLR}, initially built as \textbf{Auto Launch Review} and later extended upstream into an autonomous research-to-launch harness. AutoLR combines three system mechanisms: a \textbf{multi-expert council} that debates and adversarially reviews proposals; a \textbf{deterministic evidence-weighted exploration--exploitation selector} that allocates a limited trial budget across candidate directions and uses Council reranking; and a layered knowledge system that combines external research, production-system knowledge, and DASHEN-specific domain knowledge---such as game communities, player characteristics, and content-interaction patterns---with posterior evidence from configurations, patches, logs, failures, and offline outcomes. LLM agents perform semantic reasoning and code generation, while deterministic controllers retain authority over execution, metric extraction, guardrails, and persistent state transitions. --- ### CHAMP: Cross-domain Hybrid Architecture for Matchmaking and Prediction in Online Multi-Player Games Authors: Kai Wang, Ge Fan, Chaoyun Zhang, Yuyang Jiang, Yuze Liu Published: 2026-09-04 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.04870v1 Multiplayer Online Battle Arena (MOBA) games rely on matchmaking to maintain competitive balance. Our prior work, CUPID, framed matchmaking as an assignment re-optimization problem and showed that a single-mode win-rate predictor can meaningfully rebalance teams. However, deploying such a system across diverse player populations exposes three practical bottlenecks: most queueing players lack sufficient in-mode match history (cold start), skill distributions shift drastically across rank tiers (distribution inconsistency), and extreme skill segments are severely data-starved. We present CHAMP, a cross-domain matchmaking framework that resolves these deployment bottlenecks. To address data sparsity and cold starts, CHAMP replaces the target-mode-only player profile with a hybrid domain feature collection: a timestamp-ordered cross-mode short-term sequence whose slices are annotated with target-domain features, plus per-mode breakdowns of long-term, real-time and team statistics. We further propose the Domain-Aware Win-rate Network (DAWN): a Domain-aware Knowledge Extractor (DAKE) compiles target-mode attributes into learnable representations that feed Domain-Aware Temporal/Spatial/Permutation OmniNet Encoders (DATOE/DASOE/DAPOE), so that mode-conditioned representations and per-mode debiasing are learned jointly inside a single shared network. Online, one trained DAWN serves every supported mode, with per-mode position-satisfaction thresholds as the only mode-specific knob. Offline, DAWN achieves 67.73% win-rate prediction accuracy, outperforming all evaluated attention and sequence baselines. Online A/B tests across the entire League ladder of a large-scale MOBA game, from novice players up to the top-expert players served by Elite Mode, demonstrate consistent drops in imbalanced matches. For lower-tier players, CHAMP reduces the 5-minute kill crushing rate by up to 20.73%. --- ### From Interaction Traces to Persistent Skills: Online Evolution for Computer-Use Agents Authors: Longtao Hu, Xiao Liang, Linchao Zhu Published: 2026-09-04 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.04869v1 Computer-use agents can execute increasingly complex tasks in graphical interfaces, but their interaction experience is typically transient: procedural knowledge acquired from one rollout is not systematically retained, refined, and reused in later tasks. Existing skill libraries provide external procedural knowledge, yet their incremental value over the same agent operating without skills, as well as their longitudinal dynamics under repeated interaction, remain insufficiently characterized. We present an online skill-evolution framework that converts interaction trajectories and evaluator feedback into a persistent, versioned library of reusable procedures. Each iteration executes against a frozen library snapshot, and evidence-guided skill updates become available in subsequent iterations without changing model parameters. We compare the full evolving-library system with a configuration-matched empty-library control across four OSWorld application domains under the same fixed action-generation and GUI-grounding stack, task sets, and iteration horizons. Following a five-iteration empty-library warm-up, Full attains a higher post-warm-up mean evaluator score in all four observed domain runs, with mean differences ranging from 5.7 to 18.6 percentage points and domain-dependent temporal stability. In GIMP, provenance-aware analysis reveals retrieval across task-of-origin boundaries and revision churn, where repeated accepted edits fail to recover the originating task. These findings characterize evolving skill libraries as auditable, shared procedural memory that can improve a fixed computer-use stack, while showing that their benefits are conditional and repeated revision does not guarantee recovery. Code is released at https://github.com/LongtaoHu/Skill-Evo4GUI. --- ### PRISM-Bench: An Audio-Centric Diagnostic Benchmark for Text-to-Audio-Video Generation Authors: Yuchen Sun, Qian Yang, Jun Wang, Detai Xin, Guoqiao Yu, Guanglu Wan, Qi Jia Published: 2026-09-04 Categories: cs.MM, cs.AI, cs.SD arXiv: https://arxiv.org/abs/2609.04867v2 Text-to-audio-video (T2AV) generation has advanced rapidly, but its evaluation still underestimates the audio modality. Existing benchmarks either treat audio as an auxiliary component of video quality or assess it in isolation from audiovisual grounding, making it difficult to diagnose where current systems truly succeed or fail in audio generation. We present PRISM-Bench, the first audio-centric diagnostic benchmark for T2AV generation. Built from a rigorously curated dataset of 900 human-verified samples, PRISM-Bench factorizes audio evaluation along two orthogonal axes: audio type (Speech, Music, and Sound) and sound-source visibility (On-screen vs. Off-screen). It evaluates generated content across four perceptual dimensions (Audio-Visual Coherence, Audio Quality, Audio Expressiveness, and Prompt Following) with 35 fine-grained criteria. To ensure reliable assessment, we adopt an enhanced MLLM-as-a-Judge protocol based on blind, side-by-side comparison against ground-truth references, demonstrating strong alignment (over 70% mean agreement) with human raters. Our evaluation of recent T2AV systems highlights a significant performance gap between frontier and open-source models. Furthermore, we demonstrate that current generation paradigms overfit to perceptual fidelity while struggling with complex grounding and control tasks, particularly in generating music and synchronized On-screen audio. --- ### LLM-Assisted Behavioural and Scenario Augmentation for Agent-Based Energy Adoption Models Authors: Iias Faiud, Hossein Khaleghy, Michael Schukat, Karl Mason Published: 2026-09-04 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.04866v1 Recent advances in large language models (LLMs) create opportunities to enrich simulation-based energy policy analysis, particularly by supporting structured behavioural assumptions and exploratory techno-economic scenarios. However, directly replacing adoption models with LLM reasoning raises concerns regarding interpretability, reproducibility, and behavioural validity. This paper proposes a hybrid framework for LLM-assisted specification design, integrating bounded behavioural rubrics and structured scenario specifications into a calibrated agent-based model (ABM) of solar photovoltaic (PV) adoption by Irish dairy farms. The proposed approach preserves the original techno-economic adoption mechanism while augmenting it with bounded behavioural modulation and scenario-driven uncertainty analysis. Behavioural effects are represented through interpretable conservative, balanced, and optimistic rubrics, while future policy and market conditions are explored through fixed, rule-validated scenario specifications. Experimental results across multiple policy settings, Monte Carlo worlds, and random seeds demonstrate stable and economically plausible behaviour, with adoption outcomes remaining bounded and monotonic across behavioural regimes. The framework achieves up to approximately 13% behavioural adoption increase relative to the corresponding logistic case without producing unstable or unrealistic saturation dynamics. The results demonstrate that LLM-assisted specifications can be integrated into calibrated energy ABMs in a controlled, reproducible, and policy-relevant manner. --- ### CoSkill: Joint Reinforcement Learning of Reasoning and Meta-Skill Agents for Hierarchical Skill Evolution Authors: Jinyuan Feng, Dongmin Li, Yiqun Chen, Yang Gao, Xing Chen, Huimu Wang, Zhiqiang Pu Published: 2026-09-04 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.04865v2 Skill libraries improve the sample efficiency of agentic reinforcement learning (RL) by enabling large language model (LLM) agents to reuse procedural knowledge. Yet existing paradigms exhibit structural shortcomings: they either decouple skill evolution from policy optimization or instantiate meta-skills as fixed workflows. Both treat skills as passive objects to be managed, limiting the flexible evolution of skills and their co-adaptation with the reasoning agent. To address the limitations, we propose CoSkill, a unified multi-agent RL framework that recasts the static meta-skill workflow as a learnable Meta-Skill Agent and jointly trains it with a Reasoning Agent over a hierarchical skill library. By modeling the Reasoning and Meta-Skill Agents as a cooperative team sharing a single backbone, CoSkill enables end-to-end co-adaptation: the Reasoning Agent conditions its actions on a retrieved task skill and step skills selected from its child set, while its task performance guides the Meta-Skill Agent in refining those step skills. Experiments on ALFWorld and WebShop show that CoSkill substantially outperforms prior skill-based and RL baselines, achieving success rates of 98.4% and 90.6%, respectively (+3.5 and +6.2 pp). As shown in Figure 1, CoSkill achieves superior early-stage sample efficiency, asymptotic performance, and wall-clock efficiency. Our code is available at https://github.com/jinyuan-cookie/CoSkill. --- ### MZ-Rain: Moisture-Budget-Guided Zero-Inflated Model for Station-Level Precipitation Nowcasting Authors: Yifang Zhang, Shengwu Xiong, Henan Wang, Wenjie Yin, Yuqiang Zhang, Chen Zhou, Hua Chen, Qile Zhao, Pengfei Duan Published: 2026-09-04 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.04864v1 Accurate station-level precipitation nowcasting is critical for agriculture, water resource management, and disaster prevention, which typically is formulated as a time series forecasting problem. However, conventional time-series modeling techniques face two major challenges in addressing station-level precipitation nowcasting: (1) Lack of Physics-Guided Modeling}, where meteorological variables are treated as a homogeneous set without accounting for their distinct roles in precipitation formation, leads to predictions that deviate from the physical processes governing precipitation. (2) Severe zero inflation in precipitation, where dry intervals dominate the dataset, obscuring meaningful precipitation patterns and complicating the predictive modeling. To address these challenges, we propose \textbf{MZ-Rain}, a moisture-budget-guided zero-inflated sLSTM framework for station-level precipitation nowcasting. Guided by the moisture budget equation, MZ-Rain decomposes the precipitation formation process into process-specific pathways corresponding to moisture storage, moisture transport, surface evaporation, and precipitation persistence, and captures their temporal evolution through dedicated sLSTM branches. To account for the zero-inflated nature of precipitation, MZ-Rain introduces an adaptive Tweedie modeling strategy that adaptively modulates the rainfall mean while jointly learning precipitation occurrence as an auxiliary task, enabling the model to better balance dry-wet discrimination and quantitative precipitation estimation. Extensive experiments across diverse geographical and climatic regimes demonstrate that MZ-Rain consistently outperforms strong baselines on multiple evaluation metrics, including CSI, FAR, MSE, and MAE. In particular, the model exhibits superior skill in forecasting heavy precipitation events, while benefiting from physically grounded process modeling. ---