arXiv:2605.10522v2 Announce Type: replace Abstract: Money laundering is not only about moving illicit funds, but about hiding the money's origin and traces to complicate detection. Financial criminals resort to many methods to avoid regulators and legal thresholds. But analysts investigating alerts, dedicated to pin mule accounts and track suspicious transactions daily, also have theirs. Network visualizations can be key in countering adversarial money laundering activities, especially if they provide a clear overview of the money flows and a seamless analysis experience, but they are often not structured for this type of task. That is why we propose a tabular sequential graph visualization tailored to money laundering analysis - following transactions (edges) from the victim account that triggered an alert through multiple accounts (nodes) and banks (rows). To reduce the number of nodes and edges, we propose three methods for grouping these tabular sequential graphs: an amount-based approach, a time-based approach, and a combined solution that considers both the transaction amount and its order. A user study with experts revealed that the most effective method in node reduction was not necessarily the most interesting for analysis and that there is a trade-off between manual work and time for interpretation in more granular graphs.
Science Journals
arXiv:2605.11007v2 Announce Type: replace Abstract: We show that the core components of the Transformer -- attention, residual connections, and normalization -- arise naturally from a single geometric state estimation problem. Modeling the latent state in polar form, with direction constrained to the hypersphere and uncertainty decomposed into radial and tangential components, yields a precision-weighted filtering procedure in which normalization enforces the hyperspherical constraint, attention aggregates directional evidence, and residual connections implement incremental state updates. Under suitable first-order approximations, this estimator reduces to the standard Transformer block with rotary positional encodings, showing that its architecture follows from the underlying estimation problem rather than from independent design choices. Retaining higher-order geometric corrections yields the proposed \textit{Polar Transformer}, which more faithfully approximates the underlying radial-tangential state estimator.
arXiv:2605.18336v2 Announce Type: replace Abstract: Plasma wakefield excitation driven by two color Laguerre Gaussian laser pulses carrying orbital angular momentum is investigated analytically and through quasi-cylindrical particle in cell simulations. Using a perturbative framework together with the quasistatic approximation, the influence of the transverse laser mode structure on the longitudinal and transverse wakefields in an underdense plasma is examined in the weakly relativistic regime. The results show that drivers with finite azimuthal index produce reduced and less regular on-axis longitudinal wakefields compared to conventional Gaussian drivers. However, radial longitudinal field distributions reveal that this reduction originates from a redistribution of the wakefield energy toward finite radii rather than a simple loss of wake excitation. Orbital angular momentum carrying modes generate hollow and ring shaped wake structures accompanied by strongly modified transverse electric fields and broader plasma density perturbations. Mixed Gaussian Laguerre Gaussian configurations exhibit intermediate behavior, combining weak on-axis acceleration with pronounced off axis wake excitation. The study demonstrates that structured two-color laser drivers fundamentally modify the topology of plasma wakefields and provide an additional mechanism for controlling transverse plasma dynamics, off-axis acceleration, and angular momentum mediated wakefield structures in plasma based accelerator schemes.
arXiv:2603.22048v5 Announce Type: replace Abstract: Resolving complex code defects from natural language descriptions remains a fundamental software engineering challenge. Recently, large language models (LLMs) have driven the creation of agent-based automated repair systems. While improving repository-level problem-solving, current methods struggle with complex defects like intricate polymorphic control flows and implicit type degradation. These approaches rely on static analysis and shallow execution feedback, lacking the ability to monitor intermediate execution states. Consequently, agents often fall into speculative exploration, consuming significant tokens without identifying the root cause. We introduce DAIRA (Dynamic Analysis-enhanced Issue Resolution Agent), a pioneering automated repair framework deeply embedding dynamic analysis into the agent's decision loop. DAIRA employs a Test Tracing-Driven workflow, using lightweight tools to capture runtime evidence (e.g., call stacks and variable states) and convert it into structured semantic reports. By illuminating execution paths and causal dependencies, DAIRA enables precise fault localization and prevents context window flooding from irrelevant code retrievals. This shifts the agent's approach from speculative reasoning to deterministic inference. Evaluations on the SWE-bench Verified benchmark show DAIRA achieves a state-of-the-art 79.4% resolution rate when powered by Gemini 3 Flash Preview. Furthermore, it demonstrates robustness in addressing deep-seated logical defects, securing a 44.4% resolution rate on the most demanding tasks. Compared to baselines, DAIRA uniquely resolves complex edge cases and improves operational efficiency. Across various LLMs, it reduces inference costs by approximately 10% and input token consumption by 25%.
arXiv:2604.12144v2 Announce Type: replace Abstract: Scientific research based on multimodal clinical data (including medical imaging) requires coordinating clinical, radiological, programming, and biostatistical expertise, a fragmented process that bottlenecks discovery. We present VERITAS (Verifiable Epistemic Reasoning for Image-Derived Hypothesis Testing via Agentic Systems), a clinical co-scientist: a multi-agent system that autonomously tests natural-language hypotheses and produces a fully auditable evidence trail, tracing every conclusion through executable outputs from analysis plan to segmentation masks to statistical code to final verdict. Unlike prior AI-scientist systems, which mainly operate on tabular or text data, VERITAS grounds autonomous discovery directly in medical images. It decomposes the workflow into four phases handled by role-specialized agents, and introduces an epistemic evidence label framework that mechanically classifies outcomes as Supported, Refuted, Underpowered, or Invalid by jointly evaluating significance, effect direction, and study power. This distinction is critical in medical imaging, where non-significant results often reflect insufficient sample size rather than absent effects. We construct a tiered benchmark of 64 hypotheses spanning six complexity levels across cardiac and brain glioma MRI datasets. VERITAS reaches 81.4% verdict accuracy with frontier models and 71.2% with locally-hosted open-weight models (8-30B), outperforming all single-model baselines in both classes. It also produces the highest rate of independently verifiable statistical outputs (86.6%), so even its failures remain diagnosable through artifact inspection. Structured multi-agent decomposition thus substitutes for model scale while preserving the verifiability that scientific discovery demands. We release code, hypothesis bank, and evaluation pipeline at https://github.com/LucZot/veritas.
arXiv:2604.25421v3 Announce Type: replace Abstract: Federated fine-tuning provides a practical route to adapt large language models (LLMs) on edge devices without centralizing private data. However, in mobile deployments, the training wall-clock is often dominated by straggler-limited uplink communication under heterogeneous bandwidth, intermittent participation, and non-IID client data. Although parameter-efficient fine-tuning (PEFT) methods such as LoRA and QLoRA reduce local memory and trainable parameters, repeated transmission of adapter updates remains a major bottleneck. We propose Fed-FSTQ, a semantic-sensitivity-aware communication-control primitive for communication-efficient federated LLM fine-tuning. Fed-FSTQ uses a lightweight token-level Fisher proxy to estimate semantic sensitivity, couples token-guided sparsification with mixed-precision adapter-update quantization, and allocates higher communication fidelity to semantically load-bearing evidence while suppressing redundant transmission. The method is drop-in compatible with standard federated PEFT pipelines and requires no change to the server aggregation rule. Experiments on multilingual QA and medical QA under non-IID partitions show that Fed-FSTQ reduces cumulative uplink traffic required to reach a fixed quality threshold by 46-fold relative to a Fed-LoRA baseline and improves straggler-limited wall-clock time-to-accuracy by 52%. Under the corrected Controlled LTE-20Mbps accounting, Fed-FSTQ reduces per-round time from 414.60s to 67.29s and reduces per-round energy from 839.20J to 146.28J, yielding a 6.16-fold speedup. On NVIDIA Jetson-class edge devices, Fisher-guided token reduction also yields up to a 1.55-fold inference speedup, demonstrating deployability under tight resource constraints.
arXiv:2605.22487v2 Announce Type: replace Abstract: Recent advances in Multilingual Large Language Models (MLLMs) have significantly enhanced cross-lingual conversational capabilities, yet modeling culturally nuanced and context-dependent communication remains a critical bottleneck. Specifically, existing state-of-the-art models exhibit a severe pragmatic gap when handling structural variations, regional idioms, and honorific consistencies in low-resource contexts like Bangla. To address this limitation, we introduce a novel, culturally aligned instruction-tuning dataset for \textbf{BangLa Application and DialoguE generation - BLADE} and benchmarking framework comprising $4,196$ meticulously curated interaction pairs. We leverage this resource to systematically fine-tune and evaluate leading open-weight architectures, including DeepSeek-8B and LLaMA-3.2-3B, utilizing parameter-efficient fine-tuning via LoRA adapters in a 4-bit NormalFloat (NF4) quantization framework. Our empirical evaluations demonstrate that models fine-tuned on our dataset yield substantial improvements in structural fidelity and honorific alignment, providing a rigorous benchmark for bridging pragmatic disparities in low-resource multilingual text generation. Code and dataset: https://github.com/ashuvo25/Bangla_Application_LLM/tree/main
arXiv:2606.03003v2 Announce Type: replace Abstract: A latent world model built from an equivariant encoder and predictor inherits a provable symmetry of its training loss: when the dynamics carries a group $G$ acting on latents by an orthogonal representation $\rho(g)$, the one-step prediction relMSE is exactly invariant across the whole group, so fitting a restricted slice of orientations mathematically determines it on the entire orbit. The symmetry survives a real Muon/AdamW+EMA+VICReg run -- composed residual $\sim 10^{-6}$ after training, under any optimiser (intrinsic Vector-Neuron/e3nn parametrisation) -- and one-step error is flat across the group (5-seed medians: equivariant $\times 1.00$ vs a higher-capacity non-equivariant baseline $\times 12.7$ in 2D, $\times 17.2$ in 3D), while that baseline fits the slice but breaks out-of-distribution. The flatness is not a synthetic artefact: on real-robot DROID end-effector trajectories the equivariant model stays flat across the orbit ($\times 1.000$, rotation residual $1.5\times 10^{-16}$) while a $4.5\times$-larger baseline degrades $\times 11$. One caution is load-bearing: flatness is necessary, not sufficient -- the theorem transports the in-distribution error level unchanged but does not lower it (3D relMSE $\approx 0.43$): across-group error is constant, not low. The same isometry lifts to a closed-loop corollary: under a matching equivariant planner the control error is invariant across the group -- float-floor-exact in 2D/SO(2), statistically flat in 3D/SE(3). Stress-tested against Sutton's Bitter Lesson (augmentation, scale, soft-equivariance), each closes at most the across-group task metric, never the float-floor exactness. This is the generalisation-side foundation of a certified-world-models programme (arXiv:2606.13092, 2606.24945, 2606.24946): flatness transports competence, and the trust bounds built on it are downstream products.
arXiv:2605.25098v2 Announce Type: replace Abstract: Strong localization of light in 3D dielectric disorder is challenging to establish because localized contributions can be hidden by diffusive leakage and finite-sample loss. We perform full-wave time-domain Maxwell simulations of dense high-index dielectric slabs and analyze late times after the diffusive component escapes. Transmission develops a non-exponential tail, and an effective diffusion coefficient approaches localized-dynamics scaling. Late-time spectra show narrow resonances with sub-unity mean Thouless conductance and Poisson-like spacings. Time- and frequency-resolved near fields reveal compact non-propagating modal domains separated by persistent low-intensity channels and interacting weakly. These converging signatures constitute finite-system evidence for resolved 3D Anderson localization, with the late-time field forming a self-organized, interference-defined confinement pattern.
arXiv:2506.15857v2 Announce Type: replace-cross Abstract: Decellularized uterine extracellular matrix (dUECM) is promising for uterine tissue engineering because of its inherent bioactivity and structural complexity. However, transforming dUECM into porous, functional 3D constructs remains challenging. This study aimed to (1) synthesize dUECM using a modified decellularization protocol and formulate it into a hydrogel ink, and (2) fabricate 3D-printed constructs to support human uterine myometrial cell growth in vitro. Porcine uterine tissues were decellularized using 1% Triton X-100 with varying concentrations of sodium dodecyl sulfate (SDS) (0.1-1.5%) for 48-72 h. The resulting dUECM was characterized using DNA and glycosaminoglycan (GAG) quantification, Picrosirius Red-polarized light microscopy, histology, scanning electron microscopy, FTIR, Raman spectroscopy, and thermogravimetric analysis. To prepare the ink, dUECM powder was enzymatically digested with pepsin and blended with 2% and 3% alginate. Constructs were fabricated using extrusion-based 3D printing and assessed for filament fidelity, swelling, degradation, and mechanical properties. Biocompatibility was evaluated using hTERT-HM myometrial cells through MTT assays, Live/Dead staining, and alpha-SMA immunohistochemistry. The optimal protocol (1% Triton X-100 + 1% SDS for 48 h) reduced DNA to 51.3 +/- 9 ng/mg while retaining high GAGs (54.9 +/- 7.6 ug/mg). Preservation of the ECM structure was confirmed by spectroscopy. The 3% Alg + 1.5% dUECM hydrogel exhibited suitable printability (1.5 +/- 0.2), swelling (47 +/- 12%), degradation resistance (94 +/- 18% mass retention), and mechanical strength (323 to 175 kPa over 14 days), with high viability and proliferation (258 +/- 13%). The developed dUECM-based hydrogel supports 3D bioprinting with strong mechanical and biological performance, offering a promising platform for uterine tissue engineering.
arXiv:2604.02091v2 Announce Type: replace Abstract: Rerankers play a pivotal role in refining retrieval results for Retrieval-Augmented Generation. However, current reranking models are typically optimized on static human annotated relevance labels in isolation, decoupled from the downstream generation process. This isolation leads to a fundamental misalignment: documents identified as topically relevant by information retrieval metrics often fail to provide the actual utility required by the LLM for precise answer generation. To bridge this gap, we introduce ReRanking Preference Optimization (RRPO), a reinforcement learning framework that directly aligns reranking with the LLM's generation quality. By formulating reranking as a sequential decision-making process, RRPO optimizes for context utility using LLM feedback, thereby eliminating the need for expensive human annotations. To ensure training stability, we further introduce a reference-anchored deterministic baseline. Extensive experiments on knowledge-intensive benchmarks demonstrate that RRPO significantly outperforms strong baselines, including the powerful list-wise reranker RankZephyr. Further analysis highlights the versatility of our framework: it generalizes seamlessly to diverse readers (e.g., GPT-4o), integrates orthogonally with query expansion modules like Query2Doc, and remains robust even when trained with noisy supervisors.
arXiv:2606.05575v2 Announce Type: replace Abstract: Generative models have shown promising results for speech enhancement (SE), but they often rely on multi-step inference, limiting low-latency deployment. We propose SB-RF, a one-step generative framework that integrates Rectified Flow (RF) with Schr\"odinger Bridge (SB) theory. During training, SB-RF samples intermediate states from an SB time marginal and trains a conditional velocity field with the RF velocity-matching objective. At inference, SB-RF starts from the noisy observation and applies a single Euler update. Experiments show that SB-RF achieves competitive performance among generative methods on the VoiceBank-DEMAND benchmark. To further assess performance beyond this standard setting, we evaluate SB-RF on a simulated low signal-to-noise ratio test set using an expanded training dataset. Under these conditions, SB-RF achieves superior performance over the compared baselines, supporting its potential for real-world applications.
arXiv:2606.31536v3 Announce Type: replace Abstract: As Quantum Machine Learning (QML) transitions toward practical implementation, the field faces a critical architectural bottleneck that challenges the fundamental assumptions of classical statistical learning theory. In classical deep learning, increasing model capacity typically risks overfitting. However, this study advances a counter-intuitive paradigm: unstructured contemporary QML architectures suffer from a profound state of quantum underfitting, driven by the "expressivity-trainability paradox." We demonstrate that the vast Hilbert space capacity of Parameterized Quantum Circuits (PQCs)-traditionally chased as the source of quantum advantage is the direct mathematical cause of Barren Plateaus (BPs), where gradient landscapes become exponentially flat. By synthesizing recent breakthroughs in Dynamical Lie Algebras (DLAs) and Geometric QML, we establish a comprehensive framework linking the algebraic dimension of circuit generators to their optimization dynamics. Furthermore, we empirically validate this framework on a non-linear binary classification task, illuminating a uniquely quantum manifestation of the bias-variance tradeoff: while unstructured architectures achieve near-perfect training accuracy via unscalable parameterization (quantum overfitting), embedding group-theoretic geometric priors acts as a structural regularizer. By restricting the DLA growth to a polynomial regime, our symmetry-preserving approach sacrifices raw memorization capacity to guarantee scalable, gradient-rich training landscapes, offering a robust roadmap for "Trainability-by-Design" in scalable quantum neural networks.
arXiv:2606.31947v2 Announce Type: replace Abstract: State-of-the-art speech datasets predominantly focus on widely spoken languages, often overlooking low-resource languages such as Luxembourgish, which remain underrepresented in speech technology research. In this work, we introduce LuxEmo, a 21-hour conversational expressive speech corpus for Luxembourgish with 4 emotion categories. LuxEmo is derived from Radio T\'el\'evision Luxembourg (RTL) youth broadcasts, using automated detection followed by human validation. We propose a semi-automatic curation workflow combining voice activity detection, denoising, language identification, LuxASR-based segmentation, automatic emotion prediction, lexical cues, and targeted human review. Additionally, we benchmark five expressive TTS systems covering German-based cross-lingual transfer, multilingual Luxembourgish support, Luxembourgish adaptation, and non-parametric prosody transfer. Performance is evaluated using both objective metrics and human evaluation.
arXiv:2510.04484v2 Announce Type: replace Abstract: The ability to control LLMs' emulated emotional states and personality traits is an essential step in enabling rich, human-centered interactions in socially interactive settings. We introduce PsySET, a Psychologically-informed benchmark to evaluate LLM Steering Effectiveness and Trustworthiness across the emotion and personality domains. Our study spans four models from different LLM families paired with various steering strategies, including prompting, fine-tuning, and representation engineering. Our results indicate that prompting is consistently effective but limited in intensity control, whereas vector injections achieve finer controllability while slightly reducing output quality. Moreover, we explore the trustworthiness of steered LLMs by assessing safety, truthfulness, fairness, and ethics, highlighting potential side effects and behavioral shifts. Notably, we observe idiosyncratic effects; for instance, even a positive emotion like joy can degrade robustness to adversarial factuality, lower privacy awareness, and increase preferential bias. Meanwhile, anger predictably elevates toxicity yet strengthens leakage resistance. Our framework establishes the first holistic evaluation of emotion and personality steering, offering insights into its interpretability and reliability for socially interactive applications.
arXiv:2607.00595v2 Announce Type: replace Abstract: Gaussian Splatting has achieved significant improvements by incorporating warping-based techniques. However, such methods suffer from pixel-level inaccuracies due to uncertain geometry. This uncertainty leads to spatial misalignments in the warped images, which disrupt residual learning used in warping-based methods and fundamentally limit the gains of correction, particularly on thin structures and high-frequency details. Driven by our insight that useful visual cues are not lost but locally preserved under slight displacement, we propose Geometry-Aware Deformable Aggregation (GADA). This method introduces an iterative refinement module with deformable offsets to actively correct spatial misalignments and recover these displaced cues. Furthermore, to address the limitations of standard pipelines where visibility checks (i.e., thresholding) often discard valid pixels and multi-view warped image fusion relies on naive mean aggregation, our module is coupled with an implicit confidence weighting mechanism that selectively suppresses unreliable evidence. Consequently, our approach outperforms prior warping-based Gaussian Splatting, preserving high-frequency quality while achieving 2.13 times faster FPS.
arXiv:2607.00761v2 Announce Type: replace Abstract: Evolutionary multitasking (EMT) has shown strong capability in solving multiple optimization problems simultaneously by exploiting latent inter-task consistency, such as similarities in promising solutions or search directions. However, most existing EMT studies remain focused on objective-driven optimization, where such consistency is mainly used to accelerate convergence toward predefined optima. In this paper, we move EMT from consistency to collaborative discovery and propose a multifactorial evolutionary algorithm with collaborative discovery (MFEA-CoD) for multitask novelty search. Unlike conventional EMT, MFEA-CoD coordinates multiple novelty search tasks to collaboratively discover behaviorally novel solutions rather than merely transferring consistent search information for faster convergence. Specifically, a multitask repulsion operator encourages different tasks to explore distinct regions of the unified search space, thereby reducing redundant behavioral discoveries. Meanwhile, an adaptive inter-task transfer mechanism exploits shared discovery opportunities in overlapping novelty-improving regions by adjusting the transfer probability according to the online contribution of transferred information. Furthermore, MFEA-CoD is extended to multitask novelty-augmented optimization, where behavioral novelty is jointly considered with objective information to alleviate premature convergence caused by deceptive objectives. Experiments on synthetic basin-type problems, deceptive maze navigation problems, MuJoCo policy optimization problems, and generative novelty search problems demonstrate that MFEA-CoD improves the efficiency of discovering diverse novel solutions and shows clear advantages in deceptive objective landscapes.
arXiv:2510.16031v2 Announce Type: replace Abstract: Machine learning-based precipitation nowcasting relies on high-fidelity radar reflectivity sequences to model the short-term evolution of convective storms. However, the development of models capable of predicting extreme weather has been constrained by the coarse resolution (1-2 km) of existing public radar datasets, such as SEVIR, HKO-7, and GridRad-Severe, which smooth the fine-scale structures essential for accurate forecasting. To address this gap, we introduce Storm250-L2, a storm-centric radar dataset derived from NEXRAD Level-II and GridRad-Severe data. We algorithmically crop a fixed, high-resolution (250 m) window around GridRad-Severe storm tracks, preserve the native polar geometry, and provide temporally consistent sequences of both per-tilt sweeps and a pseudo-composite reflectivity product. The dataset comprises thousands of storm events across the continental United States, packaged in HDF5 tensors with rich context metadata and reproducible manifests.
arXiv:2510.08220v3 Announce Type: replace-cross Abstract: The performance of silicon nano-devices at cryogenic temperatures is critical for quantum qubit control circuits and space applications. Using multi-valley Monte Carlo simulations, we investigate electron transport in Si~(110) systems. At low electric fields, phonon absorption becomes negligible, and mobility is governed by competition between remote Coulomb scattering~(RCS) at low inversion charge density and surface roughness scattering~(SRS) at high density, leading to a mobility peak. High-$\kappa$ dielectrics such as $\mathrm{HfO_2}$ introduce remote phonon scattering~(RPS), which suppresses mobility. Under high electric fields, phonon emission dominates at 4~K, limiting velocity enhancement and resulting in limited current improvement
arXiv:2511.02064v5 Announce Type: replace Abstract: Grid-based methods are the most robust approach for automatic hexahedral (hex) meshing, but they struggle to achieve high boundary fidelity and element quality. Conventional pipelines remove outside elements. This yields axis-aligned surfaces that converge to the input geometry at first order. The subsequent padding and projection steps are heuristic, offering no guarantees on final boundary fidelity or mesh quality. This paper introduces MCHex, a fundamental reformulation of boundary and mesh quality handling in grid-based hex meshing. MCHex directly applies Marching Cubes (MC) inside each grid cell, ensuring that the mesh boundary is an MC surface. The key insight is that by constraining cut-cell configurations to a 3-regular polyhedron, a midpoint subdivision of these configurations produces an all-hex mesh with a guaranteed positive Jacobian for every element. MCHex provides three advantages: (1) a theoretical guarantee of positive Jacobian for all hex elements; (2) boundary convergence that matches the approximation rate of MC, together with a non-heuristic algorithm that has well-bounded time complexity and achieves the fastest wall-clock time among all existing methods; and (3) generation of manifold surfaces for arbitrary geometries and a natural padded layer. Extensive evaluation on a benchmark of 202 geometries compares MCHex against several previous state-of-the-art hex meshing methods under a slightly smaller element budget, demonstrating that MCHex consistently produces positive Jacobian meshes with similar boundary fidelity while running significantly faster. MCHex can integrate seamlessly with post-processing steps such as mesh smoothing, mesh simplification, and is suitable for simulation.
arXiv:2511.02496v2 Announce Type: replace Abstract: We study latent geometry as an explicit component of representation quality in data-scarce learning. For an encoder (\phi), we define (Q_{\beta,\gamma}(\phi)=I(\phi(X);Y)-\beta\mathcal C(\phi)-\gamma d_{\mathrm{int}}(\phi)), combining task-relevant information with penalties for curvature and intrinsic latent dimension. Thus geometry becomes part of the bottleneck criterion, not only a post hoc diagnostic. Under smooth-manifold, loss-transfer, and estimator-concentration assumptions, we derive non-asymptotic low-label generalization bounds where intrinsic dimension and covering complexity enter explicitly. We characterize the information--geometry frontier and prove empirical-surrogate consistency. The analysis links encoder geometry to learning through latent covering numbers, loss-class entropy, and uniform deviation. We instantiate the theory as \texttt{V-GIB}, adding curvature and dimension penalties to variational bottleneck training. Real low-label benchmarks compare \texttt{V-GIB} with ERM, VIB, and ablations across (1%)--(20%) label fractions. Results show improved performance and reduced geometric complexity in several regimes, especially FashionMNIST and CIFAR-10, while confirming that no fixed regularizer is universally dominant.
arXiv:2606.22982v2 Announce Type: replace Abstract: Multi-arm manipulation demands precise spatiotemporal coordination, yet many centralized approaches scale poorly as team size increases. To address this, we propose CLS-DP, a decentralized multi-agent framework that enables implicit coordination under partial observability without shared global views, explicit state information, or inter-agent communication. Under the centralized training and decentralized execution (CTDE) paradigm, CLS-DP distills privileged multi-agent dynamics into a latent space. At deployment, each agent infers a collaborative latent from its local RGB observation and a shared task instruction; it then conditions the diffusion denoising process on this latent. This design enables implicit coordination with a per-agent cost independent of team size. Across six RoboFactory benchmark tasks spanning two to four agents, CLS-DP achieves a 38% mean success rate, outperforming the best centralized baseline (20%) and a decentralized ablation without the collaborative latent (9%). It also maintains superior parameter efficiency across all agent configurations. Attribution maps show that an agent conditioned on the collaborative latent places high attribution on the joints and grippers of both itself and its teammates throughout execution. This suggests that the learned latent efficiently encodes collaborative dynamics from local observation, which facilitates implicit coordination in realistic settings characterized by partial observability.
arXiv:2606.28016v2 Announce Type: replace Abstract: Autoregressive (AR) video diffusion models enable low-latency streaming generation by synthesizing videos chunk by chunk with cached visual context, but this chunk-wise formulation makes temporal instruction following ambiguous. A single global prompt does not specify which sub-event should be realized in each chunk, while naively switching to step-wise prompts often leads to delayed reactions, blended step semantics, and error propagation across prompt transitions. These failures are difficult to address with supervised fine-tuning or distillation alone: SFT suffers from exposure bias, while rollout-based distillation still optimizes low-level denoising or teacher-distribution matching rather than directly enforcing action ordering and prompt-transition correctness. We address these challenges with TempAct, a planner--executor reinforcement learning framework that jointly optimizes temporal decomposition and step-conditioned execution for temporally plausible AR video generation. TempAct uses an LLM planner to explore span-aware step prompts that are executable by the video model, and trains an AR diffusion executor to follow these prompts under its own generated histories. Its key mechanism is hierarchical group exploration: candidate plans form planning groups, and each plan induces an execution group of multiple continuations from a shared visual context, enabling plan-level credit assignment for long-horizon temporal outcomes and executor-level credit assignment for prompt-switch behavior. We further design hierarchical rewards that combine plan-quality and full-video temporal feedback for the planner with local transition-level step-following rewards, aesthetic regularization, and KL constraints for the executor. Experiments on Self-Forcing and LongLive show that TempAct improves temporal consistency while preserving overall visual quality.
arXiv:2606.28366v2 Announce Type: replace Abstract: Shadow bands are transient, rippling patterns of light and dark that may appear in moments before and after totality in a solar eclipse. Despite centuries of reports, their physical origin has remained unresolved. This preprint develops a geometric-optical solution in which the Sun's extended structure produces a celestial analogue of Young's double-slit experiment, generating an interference-like intensity pattern on the ground, modulated by Earth-atmosphere effects. The analysis combines solar limb geometry, atmospheric propagation effects, and a wave-based formulation that yields quantitative predictions for fringe width and spacing. The resulting model accounts for the principal observational features of shadow bands and clarifies why the phenomenon is both elusive and highly sensitive to viewing conditions.
arXiv:2606.28864v2 Announce Type: replace Abstract: Test-time scaling is a paradigm where large models use additional compute at inference to achieve better performance, without changing model weights. While it has been widely studied for Large Language Models (LLMs), its applicability to Large Vision-Language Models (LVLMs) remains less explored and analyzed, with limited analysis of whether, when, and to what extent these approaches transfer to LVLMs. In this work, we ask a simple but fundamental question: can conventional test-time scaling methods developed for LLMs be directly applied to LVLMs? We present the first comprehensive study of test-time scaling for LVLMs, spanning multiple models and model sizes, nine test-time scaling methods, and six diverse benchmarks. Our main findings is that 1) different from previous findings, small, well-performing models benefit the most from test-time scaling, enabling performance improvements of up to around 30\%, reaching large models performance, and often outperforming them, 2) LVLMs lose focus when given more compute than necessary, and 3) Visual information is encoded early in the reasoning chain, after which the chain is dominated by text-only reasoning and the contribution of image tokens drops significantly. Finally, we also provide a global and fine-grained analysis on the quality and information sufficiency of the reasoning chains produced. Overall, our findings and analysis provide practical guidance and insights into LVLMs and their deployment in research and industry.