Papers

Fallacy-detection benchmarks pair fallacy classes with a single "valid" or "none" class that takes everything data collection did not label as a fallacy. A detector has two jobs, deciding whether an argument is fallacious and naming which fallacy it commits, and the false-positive rate is meant to measure the first. We show that what these benchmarks actually score is scheme recognition, the ability behind the second job. Their own test sets already show it: when a classifier misses a fallacy, the error lands on "none" rather than on another fallacy type, so detection is failing while classification holds. The reason is what the valid class lacks. The negatives that separate the two jobs are correct arguments using the same argumentation scheme as a fallacy, and they are scarce: nearly absent from the four benchmarks we examined, and rare even under deliberate search. A detector is therefore never tested where recognizing a scheme and judging its use come apart, and can pass on recognition alone. We construct the missing arguments, together with a control condition from the same pipeline that differs only in scheme, so whatever generation contributes, it contributes to both. The classifier labels the scheme-matched negatives as the source fallacy, and labels the wrong-scheme negatives as the scheme they actually use 85.9% of the time and as the source type 0.4%. The classifier has learned which scheme an argument uses, not whether it uses it correctly. The over-flagging follows: a model that scores 16.6% on CoCoLoFa's own valid class flags 58.9% of the constructed arguments. The same dissociation appears in three zero-shot LLM detectors that never saw these benchmarks. We release the items as Scheme Foils. A reported false-positive rate should not be trusted as a measure of detection until the valid class has been audited for scheme-matched coverage.

In this paper, we derive a new lower bound for the Hausdorff dimension of 3D Brownian cut points by proving an explicit upper bound ($<0.9999$) for $\xi_3(1,1)$, the intersection exponent for two independent Brownian motions in 3D. Consequently, the growth exponent of 3D loop-erased random walk is at least $1.0001$.

Large language models (LLMs) often suffer from capability stagnation in self-improvement training because fixed difficulty levels fail to adapt to their evolving proficiency. To address this issue, we propose STRETCH (Self-Taught Reasoning Evolution via Targeted CHallenge), a unified framework inspired by cognitive scaffolding theory. STRETCH introduces a dynamic Stretch Zone mechanism that continuously aligns question difficulty with the model's solving capability. Within a single parameter space, the model alternates between a Scaffolder that generates adaptive, boundary-pushing challenges and a Learner that that optimizes its solving trajectories through reinforcement learning. This dual-loop co-evolution effectively stabilizes training, mitigates reward hacking and promote progressive reasoning growth. Experiments on both negotiation and operation research benchmarks demonstrate that STRETCH consistently outperforms strong prompting and domain-specific baselines. Further scaffolder configuration analysis shows that dynamic difficulty alignment is critical for sustained capability improvement and synchronized reasoning evolution.

In this manuscript, we analyze the dynamics of a model in which two spin-1/2 systems are each coupled noncommutingly to a static environmental field and to a self-Hamiltonian, while also being mutually coupled through a $\sigma_y^{(1)}\sigma_y^{(2)}$ interaction. By constructing an exact solution that exploits parity symmetry, and by applying the stationary-phase approximation (the saddle-point method) in the continuum limit of the environmental field, we show that coherence associated with the ordinary isolated stationary point decays as $t^{-1}$, whereas, under the resonance condition at which the local effective fields of the two spins cancel each other, only the coherence between a pair of maximally entangled, Bell-type states- the eigenbasis of the inter-system coupling-survives, decaying via the anomalously slow power law $t^{-1/2}$. Through a comparison with the pointer-basis theory of W. H. Zurek and coworkers (the 1981 and 2005 models), we further reveal a mechanism not reducible to local coupling: in this system, the pointer observable is determined not by the individual system-environment couplings but by a statistical resonance condition between the two environments.

Integrated Kerr microcombs have emerged as a central platform for chip-scale nonlinear photonics. One of their most compelling applications is dual-comb technology, which enables fast, high-resolution, coherent optical spectroscopy, precision metrology, ranging, and ultrasensitive sensing by interfering two frequency combs with slightly different repetition rates. However, no effective strategy yet exists to synergistically excite multiple spatial mode families, manage mode competition and energy partitioning, and precisely engineer dispersion, leaving offset co-propagating dual-microcombs in a single resonator elusive. Here we demonstrate a new mechanism for offset dual microcombs formation via Kerr-Raman interleaving for the first time, by leveraging single Raman-active thin-film lithium niobate microrings. Two spatial mode families with high loaded Q factors up to 5 million are dispersion engineered to have nearly identical free-spectral-ranges but a designed relative frequency offset. Stimulated Raman scattering efficiently generates a Raman comb in one family alongside a Kerr comb in the other. The two combs interleave with a fractional offset to form a broadband, densely interleaved spectrum. Pumping the microring at only 320 mW yields Kerr-Raman interleaved dual combs both featuring nearly identical repetition rates around 102 GHz with a small difference of 675 MHz, a relatively flat envelope exceeding 29 nm and an intrinsic comb-line linewidth of 410 Hz. Our results establish offset dual-comb generation in a single microresonator, enabling compact dual-comb sources for chip-scale precision spectroscopy, Vernier-enhanced sensing, and advanced quantum photonics.

Emergency management assistance programs, such as relief distribution, are essential for delivering necessary supplies to affected communities. However, these programs operate in a decentralized network of local centers that face uncertain local demand and supply dynamics, resulting in inconsistent avail- ability of local services. Redistribution of supplies among these local centers reduces these imbalances, but the centers often make decisions independently, with limited information and disrupted transportation. This study develops CoRe-MARL, a cooperative multi-agent reinforcement learning (MARL) framework, by formulating a decentralized partially observable Markov decision process (Dec-POMDP). We treat each center as an agent that learns a redistribution policy to improve the service in the worst-case region and reduce the service gap across regions while protecting network-wide service. We incorporate a recurrent network that captures evolving supply and demand dynamics without direct observation, while multi-agent proximal policy optimization (MAPPO) enables centralized training and decentralized execution (CTDE). We evaluate the framework in a simulated environment with diverse trajectories, where exact dynamics are not observed by actors and the MAPPO critic. We compare the recurrent MAPPO with the recurrent independent PPO (IPPO) and a local only heuristic, and find that MAPPO reduces the service gap across local centers and enhances service for the worst-served center while maintaining competitive network-wide service. The recurrent MAPPO also shows consistent performance across diverse trajectory patterns, demonstrating its ability to adapt to evolving dynamics. The findings demonstrate the capability of cooperative learning for decentralized redistribution and improving equitable service under uncertain and evolving dynamics.

The main contribution of this work is to identify model parameters for an Anion Exchange Membrane (AEM) system directly from measured operational data, enabling their use in power system studies. Due to the relatively low technology readiness level (TRL) of AEM electrolysers, only limited literature reports operational parameters supported by openly available experimental data. By capturing the realistic dynamic behaviour and power consumption of the AEM electrolyser, this model allows for more accurate evaluation of system topologies and control strategies for green hydrogen production. This paper presents a model of an AEM electrolyser based on experimental data from Enapter modules installed in the Power-to-X (PtX) Laboratory at the Technical University of Denmark [1]. The model comprises a polarisation curve and a model of current dynamics during operation. It is validated against operational data and can be applied in grid-balancing studies as well as in assessments of how electrolyser dynamics influence the power system. The results suggest that the AEM technology is suitable for services such as smoothing wind farm power output and provide system services to support frequency balancing. The findings highlight the importance of considering the constraints of the process when designing for flexible load operation, and consideration of the interaction between the power control and process control systems.

Ground based observatories are experiencing significant technological advances and telescopes are getting equipped with highly complex instruments to enable cutting-edge astronomical research. The increasing complexity of such instruments displays a challenge to their users when they want to plan and prepare specific observation programmes, that often request to define large amount of detailed instrument-specific information. Therefore, at ESO we developed an Observation Preparation tool (in short ObsPrep tool), which provides a layer of abstraction over the instrument-specific technical details and tries to harmonize the users experience when being faced with the preparation of observation blocks for different complex instruments. The ObsPrep tool allows interactive planning of the observations by visualization of various actions and observing strategies. In a joint approach by software engineers and astronomers, ObsPrep was started as a stand-alone java desktop application and over the years has evolved into an angular based web-application. It is now fully integrated in ESOs general use phase2 user interface (p2) that allows ESO users to create, modify, or delete observation blocks. Here we present the core functionalities of ObsPrep, as well as its special features and how the user interface integrates into ESOs generic phase2 observation preparation tool. We also outline the upcoming plans to incorporate the observation preparation for the instruments of the next generation of ESOs large telescopes, the ELT, into ObsPrep.

The reuse of metal powders in directed energy deposition using laser powder offers promising sustainability benefits for additive manufacturing, yet its impact on part quality remains unknown. This study investigates the influence of powder reuse on the directed energy deposition process of Inconel 625. Virgin powder was first characterized in terms of flowability, morphology, size and chemical composition, then used to manufacture parts under fixed process parameters. These parts were analysed through two configurations (as-built and hot isostatically pressed) to assess microstructural evolution and mechanical performance. Results show that hot isostatic pressing reduces porosity and promotes grain recrystallization, enhancing ductility while decreasing tensile strength. Recovered powders from one manufacturing cycle, were also characterized. Particle size distribution and morphology remained stable, interstitial contents, particularly oxygen, increased, potentially affecting laser-material interaction and melt pool behaviour in future production. These findings highlight the importance of evaluating powder reuse through a full-process approach, from powder to final part. Future work will focus on the tensile and fatigue performance of parts produced with recovered powders.

We study the canonical Maximum Clique and Maximum Independent Set problems in the one-pass edge-arrival graph streaming setting. Here, the edges of some input graph $G = (V,E)$ are presented one at a time (possibly including deletions), before an algorithm needs to produce either a large clique or independent set at the end of the stream, with the focus being on space complexity. We are interested in finding $\beta$-approximate solutions, for any $\beta \geq 1$. Previous work gave an algorithm using $\tilde{O}\left(n^2/\beta^2\right)$ bits of space, together with a corresponding $\tilde{\Omega}\left(n^2/\beta^2\right)$ two-party communication lower bound [Halldórsson et al., ICALP'12], seeming to resolve the problem. However, their algorithm crucially relies on randomness, and the best known deterministic algorithm remains a folklore derandomisation using $O\left(n^2/\beta\right)$ bits of space, leaving a (deterministic) gap of size $\tilde{O}(\beta)$. We make progress towards this gap with a strong lower bound for a restricted range of $\beta$: whenever $\beta \leq \sqrt{n}/36\log n$, any deterministic algorithm for either problem must use $\Omega\left(\frac{n^2}{\beta\cdot\log n}\right)$ bits of space. Our proof is via a two-party one-way communication lower bound, and highlights the power of randomness when approaching either of these problems.

Video streaming dominates global internet traffic, yet conventional pipelines remain inefficient for structured, human-centric content such as sports, performance, or interactive media. Standard codecs re-encode entire frames, foreground and background alike, treating all pixels uniformly and ignoring the semantic structure of the scene. This leads to significant bandwidth waste, particularly in scenarios where backgrounds are static and motion is constrained to a few salient actors. We introduce GenStream, a semantic streaming framework that replaces dense video frames with compact, structured metadata. Instead of transmitting pixels, GenStream encodes each scene as a combination of skeletal keypoints, camera viewpoint parameters, and a static 3D background model. These elements are transmitted to the client, where a generative model reconstructs photorealistic human figures and composites them into the 3D scene from the original viewpoint. This paradigm enables extreme compression, achieving over 99.9% bandwidth reduction compared to HEVC for the continuous data stream. We partially validate GenStream on Olympic figure skating footage and demonstrate potential for high perceptual fidelity under minimal data. While acknowledging the significant computational costs shifted to the client and challenges in generalization, GenStream opens new directions in volumetric avatar synthesis, canonical 3D actor fusion across views, and personalized viewing experiences, laying the groundwork for scalable, intelligent streaming in the post-codec era.

Cloud-native microservices architectures rely on network namespaces for isolation, with the overhead of container communications remaining a critical performance bottleneck. While colocating containers on the same host mitigates some of this overhead, it cannot match the performance of communication within a single network namespace. Existing solutions either require application rewrites or fail to support the full Linux network stack expected by containerized applications. In this paper, we present netkit, an eBPF-based datapath that specializes the Linux networking stack to eliminate redundant backlog queue traversals during network namespace transitions. netkit leverages eBPF to transparently redirect packets between namespaces, bypassing unnecessary buffering while preserving compatibility with existing container applications. Our implementation in the Linux kernel, integrated with minimal changes to the Cilium network plugin for Kubernetes, improves throughput by up to 37% and achieves parity between container-to-container and process-to-process communications, effectively closing the performance gap introduced by namespace isolation.

Multi-modal talking avatar synthesis aims to generate realistic talking videos from a reference portrait and speech. Despite rapid progress in diffusion-based methods, existing approaches still struggle to jointly achieve accurate lip articulation, human-preferred motion aesthetics, and efficient inference. We observe that phonetic accuracy and motion aesthetics arise from fundamentally different sources and should be addressed at complementary stages rather than learned implicitly by a single generator. Based on this insight, we propose VibeAvatar, which disentangles these two objectives through a Phonetic Kinematics Adapter (PKA) that converts recognition-oriented speech features into phonetic-kinematic conditions at the conditioning stage, and an Aesthetic Motion Policy (AMP) that optimizes a flow-consistent stochastic sampling policy via Group Relative Policy Optimization (GRPO) at the post-training stage. With a lightweight flow-based motion generator operating in a compact 1D warp-based latent motion space, VibeAvatar achieves state-of-the-art results in articulation, aesthetics, and efficiency on both objective metrics and user studies, while generating a 10-second 512px video in under 10 seconds with only $\sim$3GB VRAM.

Differential embedding methods aim to reconstruct attractors of dynamical systems from noisy measured time series, but require accurate estimates of signal derivatives. We introduce SHADED (Savitzky-Golay and Homogeneous-differentiator based Automatic DEnoising and Differentiation), a novel methodology for denoising and estimation of derivatives up to an arbitrary order, which enables attractor reconstruction via differential embedding from noisy time series data. Homogeneous Differentiators (HD) guarantee finite-time derivative estimates in the presence of noise, while subsequent Savitzky-Golay (SG) filtering attenuates chattering. Crucially, SHADED extracts all parameters required for application of both HD and SG automatically from the data, without requiring manual tuning that may lead to inaccurate reconstruction, and can also incorporate prior knowledge, if available, thereby yielding a flexible tool for data-driven numerical differentiation of noisy signals. The obtained differential embeddings can reveal features of the underlying dynamics that are useful, e.g., for system identification, pattern recognition and discrimination between dynamic regimes; the latter application is particularly important in biomedical settings, to help distinguish between different physiological and pathological states. We demonstrate the efficacy of SHADED by testing it on computational neuroscience models, LTspice-simulated chaotic electronic circuits, and photoplethysmography and arterial blood pressure experimental recordings: across all these case studies, SHADED produces accurate derivative estimates and accurate attractor reconstructions via differential embedding (whenever a ground truth is available) or geometrically coherent and reproducible reconstructions consistent with the expected dynamics (in the absence of a ground truth), without the need for manual parameter tuning.

Let $G=D_{\langle2\rangle}([0,1])$ be equipped with either of its two Dlab orders. There exists an order automorphism $\alpha$ of $G$ such that, for every rank-one subgroup $H\leq\mathbb R_{>0}^{\times}$, every one of the six corresponding Dlab groups $A$ listed below, equipped with any of its linear orders, every order-preserving embedding $e:G\hookrightarrow A$, and every $u\in A$, there exists $f\in G$ such that $e(\alpha(f))\neq u^{-1}e(f)u$.

The Hajnal-Szemerédi theorem guarantees an equitable $(r + 1)$-coloring of every graph of maximum degree at most $r$. We prove a version of this result for more general prescribed color class sizes, subject only to a natural upper bound.

Visual-inertial odometry (VIO) is a core capability for autonomous operation in GPS-denied subterranean environments, yet its reliability can degrade sharply under sensor drift, calibration errors, and dynamic occlusion. Existing evaluations mainly emphasize nominal-condition accuracy, offering limited insight into when practical deployment failures occur. In this work, we present a failure-centric stress-test benchmark for VIO in underground environments using the CERBERUS dataset. We systematically evaluate four representative VIO systems spanning filtering-, optimization-, and learning-based paradigms under nine practical perturbation settings, including IMU bias and noise variation, camera intrinsic and extrinsic drift, and dynamic scene occlusion. Beyond conventional trajectory error, we analyze robustness limits through coverage ratio and failure thresholds, revealing breakdown behaviors that are not captured by nominal-condition performance alone. Our study shows distinct vulnerability patterns across VIO paradigms: some methods are more sensitive to inertial degradation, while others are more affected by geometric miscalibration or dynamic interference. These results provide deployment-oriented guidance for VIO selection, calibration prioritization, and reliable operation in challenging underground scenarios. To support reproducible evaluation and future extensions, we will release the full benchmark scripts and evaluation pipeline.

A parton density depends on the renormalization scheme, and any two schemes are connected by a transformation that preserves the DGLAP form of the evolution. In Mellin space such a transformation acts on the evolution kernels as a gauge transformation, and we take this point of view throughout the discussion. For the unpolarized densities we first ask which transformations also preserve the charge-conjugation and flavor symmetries of the evolution kernels and the momentum and valence number sum rules. The symmetries reduce the transformation to a few blocks, and the sum rules constrain only two Mellin moments, where they leave a single freedom, the split of momentum between the quark singlet and the gluon, and freeze the valence numbers. The resulting classification is complete, because with asymptotic freedom as a boundary condition a change of scheme is determined by its action on the kernels. Inside this group, the transformations that preserve positivity form a semigroup, and the same argument covers the polarized densities. A transformation that fails only the sum rules is not lost either, since a normalization that the sum rules themselves determine repairs it, and we use this to define collinear densities as integrals of transverse-momentum-dependent densities without giving up the sum rules. We then use the same gauge freedom to solve the evolution. Gauging the kernel away reduces DGLAP at any order to a single second-order linear equation plus a quadrature, and it organizes the estimate of missing higher orders in a way that respects the sum rules exactly.

Let $(A,\mathfrak{m}), (B,\mathfrak{n}) $ be Gorenstein local rings and let $\underline{CM}(A)$ be its stable category of finitely generated maximal Cohen-Macaulay $A$-modules. Suppose we have an equivalence $\Phi \colon\underline{CM}(A) \rightarrow \underline{CM}(A)$ as triangulated categories. We show (1) If $A, B$ are not hypersurfaces then $dim A = dim B$. (2) If $M$ is a maximal \CM \ $A$-module then $$\text{curv}_A(M) = \text{curv}_B (\Phi(M)),$$ here $\text{curv}_A(M) = \limsup_n \sqrt[n]{\ell(\text{Tor}^A_n(M,k))}$. (3) $A$ satisfies Serre's condition $R_i$ if and only if $B$ satisfies $R_i$. (4) $A$ is a complete intersection on the punctured spectrum of $A$ if and only if $B$ is a complete intersection on the punctured spectrum of $B$. We also show that $\Phi$ imposes constraints of residue field of $B$ in terms of residue field of $A$ and vice-versa. Finally if $I$ is an ideal in $A$ such that the extended Rees algebra $\mathcal{R}(I) = A[It, t^{-1}]$ is Gorenstein then we construct a triangulated functor $\Psi \colon \underline{CM}^\mathbb{Z}(\mathcal{R}(I)) \rightarrow \underline{CM}(A)$ where $\underline{CM}^\mathbb{Z}(\mathcal{R}(I))$ is the stable category of all graded maximal Cohen-Macaulay $\mathcal{R}(I)$-modules. We show that $\Psi$ induces an equivalence $\underline{CM}^\mathbb{Z}(\mathcal{R}(I))/\ker \Psi \rightarrow \underline{CM}(A)$. We give some applications of this map.

In this paper, we study the nonlinear Schrödinger equation with a nonlinearity combining a focusing logarithmic term and a smooth $L^2$-subcritical function. Our main result is the construction of multisoliton solutions for this equation. These special configurations behave asymptotically, for large times, as the superposition of several solitons moving with distinct velocities. The proof relies on an iterative construction method combined with localized energy estimates adapted to the mixed structure of the nonlinearity.

3D Gaussian Splatting (3DGS) enables photorealistic novel view synthesis, but transmitting gigabyte-scale scene data remains challenging for immersive applications. Traditional HTTP Adaptive Streaming over TCP introduces Head-of-Line (HOL) blocking and coarse segmenting ill-suited to fine-grained 3DGS delivery. We propose MoQSplat, which maps 3DGS content onto the Media over QUIC (MoQ) transport hierarchy. MoQSplat partitions scenes into spatial Tracks, clusters splats into semantically coherent Groups, and constructs progressive-quality Subgroups mapped to independent QUIC streams to eliminate connection-level HOL blocking. Using a stateless, subscriber-driven adaptation loop, clients dynamically request spatial regions and quality tiers based on six degrees of freedom (6-DoF) frustum visibility, distance, and foveal alignment. We evaluate the core components on a prototype implementation, showing that opacity-based pruning outperforms scale-based pruning for progressive delivery. The source code is available at https://github.com/emanuele-artioli/MoQSplat.

State-of-the-art vision-language-action models (VLA) for autonomous driving face critical limitations: excessive parameter counts, inefficient high-resolution image processing, and lack of temporal memory. We introduce Fast and EffectIVE VLA (FIVE-VLA) to address these through two key contributions. First, we employ an efficient vision encoder that processes high-resolution ($448 \times 896$) images while generating only 98 tokens, over $5\times$ fewer than existing approaches, and bypass text generation entirely for single-pass trajectory prediction. Second, we propose Recurrent Action Memory (RAM), a lightweight module that conditions action prediction on previous action tokens, providing temporal context critical for manoeuvres such as overtaking and emergency braking. With only 641M parameters, FIVE-VLA completes $\sim$10% more routes without traffic rule infractions than the previous state-of-the-art VLA on the challenging Bench2Drive closed-loop driving benchmark. Non-reactive open-loop simulation on the large-scale real-world NVIDIA Physical AI AV dataset shows 10.2% and 7.7% lower collision-violation rates than SimLingo in single- and four-view settings, respectively. Additionally, FIVE-VLA runs at $\sim$30 fps on an A100 and $\sim$4 fps on a T4 GPU (proxy to an edge device), representing an 8-30$\times$ speedup over previous methods.

Classifiers can make identical predictions yet require labels to compare their selective performance: confidence ranks weight the same errors differently. We quantify this requirement for the area under the generalized risk-coverage curve (AUGRC). A prelabel lower bound rules out insufficient budgets. With all labels known, a covering linear program bounds the minimum number of labels sufficient to fix the winner (the certificate size) within $K-1$ labels for $K$ candidates. For fixed $K$, independent uniform orders and identical predictions, the prelabel bound approaches one quarter of the pool. With iid Bernoulli errors independent of the orders, every exact acquisition policy reads almost all labels asymptotically, although a two-candidate certificate needs only half. Across 108 feature-panel comparisons on nine datasets, disagreement labels settle every accuracy choice but no AUGRC choice. A 20% budget is ruled out in 96 conditions; certificates need 56-57% on average. On ten conditions with pretrained image classifiers, confidence-score choice reads 68-91% of 10,000 labels for exact selection and 50-67% with AUGRC tolerance $5\times10^{-4}$. An exact stopping test works with any acquisition order. Together, these results link confidence ranks to label budgets and certified model comparison.

Let $\mathcal R$ be a normalized root system in $\mathbb R^N$ with a nonnegative multiplicity function $k$, and let $\mathcal F$ be the associated Dunkl transform. We prove a Hörmander multiplier theorem on the Hardy spaces $H^p_{\mathrm{Dunkl}}$, $0<p\le1$, defined by conical Littlewood–Paley square functions. Let $W_2^s$ denote the classical Sobolev space in $\mathbb{R}^N$. More precisely, if a multiplier $m$ satisfies \[ \sup_{t>0} \|\psi(\cdot)m(t\cdot)\|_{W_2^s}<\infty \] for a nonzero radial cutoff $\psi\in C_c^\infty(\mathbb R^N\setminus\{0\})$ and \[ s>\mathbf N\left(\frac1p-\frac12\right), \] where $\mathbf N$ is the homogeneous dimension, then the Dunkl multiplier \[ \mathcal T_mf=\mathcal F^{-1}(m\mathcal Ff) \] extends to a bounded operator on $H^p_{\mathrm{Dunkl}}$.

Creating deformable assets for robot manipulation requires jointly specifying their geometry, appearance, and physical properties. This is especially challenging for deformable objects, since text and images provide limited evidence about how they deform and respond to contact, yet these responses directly affect their suitability for interaction. Automated generation therefore needs to resolve coupled physical requirements and use interaction evidence to guide construction and refinement. We present DeformSmith, a framework that enables automated generation of interactive, physically credible deformable assets from text or a single image. Through hierarchical agentic construction and a shared physics-grounded harness, it progressively builds, tests, and refines geometry, physical models, material behavior, and robot interaction until the resulting asset is ready for simulation and manipulation. Robot interaction closes the generation loop through manipulation feedback and replayable interaction data. Results show that DeformSmith generates assets with better visual quality and physical plausibility than state-of-the-art baselines, including PhysGen3D, PhysGM, and PhysX-Omni, while supporting the synthesis of data for robotic manipulation of deformable objects. Project page: https://can-lee.github.io/deformsmith-web/

Let $E$ be a Hilbert C*-module and $\mathcal{L}(E)$ the C*-algebra of all bounded adjointable operators on $E$. An operator $T \in \mathcal{L}(E)$ is a Kasparov cycle if $T^*T - 1$ and $TT^*-1$ are compact operators on $E$. If $\mathcal{L}(E)\rightarrow \mathcal{L}(E)$ is a prime C*-algebra, and $\varphi:\mathcal{L}(E)\rightarrow \mathcal{L}(E)$ is a linear map which is unital and surjective up to compact operators, and preserves Kasparov cycles in both directions, then the induced map $\psi:\mathcal{L}(E)/\mathcal{K}(E) \rightarrow \mathcal{L}(E)/\mathcal{K}(E)$ is either a $*$-automorphism or a $*$-anti-automorphism. Our work extends the main result of [J. Math. Anal. Appl. 354 (2009), 625-629] to the set of Kasparov cycles, showing that the assumption “$\mathcal{L}(E)/\mathcal{K}(E)$ has real rank zero” is redundant in our results.

The present article investigates the well–posedness of a discrete coagulation–annihilation model originally introduced by E. Ben-Naim and P. Krapivsky in Physical Review E (1995). We first establish the existence of solutions for a broad class of coagulation and annihilation rates, including rates that may be unbounded. These results are obtained by combining uniform estimates with Helly's selection theorem and the Rellich–Kondrachov compact embedding theorem. Under appropriate structural assumptions on these rates, we further prove uniqueness and continuous dependence of solutions on the initial data, thereby establishing well–posedness of the model. Finally, for a specific class of coagulation and annihilation rates, we additionally study the differentiability of solutions and characterize their long–time asymptotic behaviour. These results provide a rigorous mathematical framework for the analysis of the model and contribute to a deeper understanding of its well–posedness, regularity, and long-time dynamics.

This paper studies (co)products and additive structures under Gabriel–Zisman localization $Q:\mathcal A\longrightarrow\mathcal A[S^{-1}]$. Examples show that localization may destroy existing (co)products or fail to preserve (co)products that remain, even when $Q$ is additive. For a set-indexed family $(\mathcal A_i,S_i)_{i\in I}$, necessary and sufficient conditions are given for the canonical functor $\big(\prod_{i\in I}\mathcal A_i\big)\big[\big(\prod_{i\in I}S_i\big)^{-1}\big]\longrightarrow\prod_{i\in I}\mathcal A_i[S_i^{-1}]$ to be full or faithful. These criteria yield compatibility with $I$-indexed (co)products for localizations admitting a calculus of fractions or arising from model categories, provided the relevant (co)products exist and the designated classes of morphisms are closed under them.

Subspace methods such as multiple signal classification (MUSIC) achieve super-resolution direction of arrival (DoA) estimation by exploiting the orthogonality between the array manifold and the noise subspace of the measurements. Their accuracy therefore depends on the assumed manifold and degrades under model mismatch, while parameters not identifiable from the spatial manifold cannot be recovered. In this work, we propose the conditional neural manifold (CNM), which replaces the fixed manifold with an observation-conditioned mapping from source parameters to steering vectors. An encoder maps the snapshots to a latent scene representation that conditions a zero-initialized neural field over the parameter space. The manifold is learned without steering-vector supervision by shaping the resulting MUSIC landscape. Since the correction acts on the manifold rather than on the estimator, it can be used by other manifold-based methods without modification. The CNM restores resolution under array imperfections, colored noise, correlated sources, and near-field propagation, and resolves the angle-frequency ambiguity inherent to the nominal spatial manifold.

We study the optical properties of electrically charged black holes sourced by generalized Heisenberg-Euler nonlinear electrodynamics (NLED) in the second-order formalism. Using the exact parametric form of the static spherically symmetric solution, we analyze the photon sphere, the black hole shadow, and the light deflection angle, and compare the results with those of the Reissner-Nordström and Schwarzschild cases. We find that the photon sphere and shadow radii decrease with increasing charge and increase with black hole mass, while NLED effects become significant mainly near extremality. By confronting the dimensionless shadow radius with the Event Horizon Telescope (EHT) bounds from Sgr A* and M87*, we derive phenomenological upper limits on the charge, with Sgr A* providing the stronger constraint. We further show that, for fixed impact parameter and charge, the deflection angle is systematically larger than in the corresponding Reissner-Nordström spacetime, and that the critical impact parameter is shifted to larger values. Our results show that NLED corrections produce observable modifications in the optical properties of charged black holes, especially close to extremality.