arXiv:2607.14191v1 Announce Type: new
Abstract: Pediatric electronic health records capture developmentally structured clinical trajectories, yet their potential for generative healthcare foundation models remains largely unexplored. Here we present TEDDY (Temporal Event Decoder for Disease in Youth), a 1.84-million-parameter decoder transformer trained on approximately 73 million ICD-10 diagnoses from 1.6 million children at a single pediatric institution. TEDDY models longitudinal diagnosis trajectories and visit timing. Predictions were made before visit codes were revealed, limited to first occurrences, and evaluated against sex- and age-matched controls. Across 797 disease-onset prediction tasks spanning 16 ICD-10 chapters, TEDDY achieved a median AUC of 72.0%, outperforming same-data DenseNet (50.0%), CNN (57.2%), RNN (60.1%), and LSTM (62.7%) baselines on 96-99% of tasks. Performance held across sex and age and was strongest among lower-prevalence diagnoses; 202 of the 225 rarest conditions (90%) had 95% confidence intervals above chance. Predictive signal remained detectable more than two years before first recorded diagnosis, with median AUCs of 59.7% in the unrestricted analysis and 64.4% in a fixed-cohort sensitivity analysis. In asthma and attention-deficit/hyperactivity disorder benchmarks, AUCs were 79.3% and 84.7%, compared with 62.7% and 71.7% for the strongest comparators, including a general-purpose language model three orders of magnitude larger. Visit-timing predictions had a 3.0-day mean absolute restricted mean survival-time error over 365 days, although median and long-tail return intervals remained miscalibrated. Together, these results establish pediatric diagnostic histories as a substrate for compact generative models supporting broad, rare-disease, and long-horizon risk forecasting without population-scale data or billion-parameter models.
Science Journals
arXiv:2509.19869v2 Announce Type: replace
Abstract: Data-driven control increasingly relies on deep models for complex systems whose first-principles models are difficult to obtain. For reliable deployment, however, learned dynamics should respect physical structure and lead to tractable optimal control. We introduce sign constraints, namely sign restrictions on Jacobian entries, as a unified description of monotonicity, positivity, and sign-definiteness. For exactly linearizable deep dynamics, we provide structural conditions and neural-network parameterizations that enforce these constraints by construction. The same structure also allows model predictive control to be formulated as a convex quadratic program or as a convex relaxation, yielding a unique optimizer and a Lipschitz continuous control law. Applications to a three-tank system and a hybrid powertrain demonstrate that the proposed approach offers improved extrapolation performance and smoother control inputs compared with competing nonconvex formulations.
arXiv:2512.06364v4 Announce Type: replace
Abstract: Current mobile health platforms are predominantly individual-centric and lack the support for coordinated, auditable multi-actor workflows. However, in many settings worldwide, health decisions are enacted through multi-actor coordination rather than individual users. We present JEEVHITAA, a cross-platform mobile system enabling role-aware sharing and verifiable information flows within permissioned care circles. JEEVHITAA ingests platform and device data, builds layered profiles from sensors and tiered onboarding, and enforces fine-grained, time-bounded access control across care graphs. Data stays secure both within the application and the cloud. Integrated retrieval-augmented Large Language Models produce structured, role-targeted summaries and action plans, offer evidence-grounded verification with provenance and confidence scores, and support advanced insights on reports. We describe the architecture, connector abstractions, and security primitives, and report robustness evaluations using synthetic, ontology-driven data and findings from a feasibility study with real-life care circles across 9-14 weeks. We outline plans for larger multi-site evaluations focusing on operational alignment, longitudinal trust & literacy impact, and relational friction & efforts to sink into the daily infrastructure.
arXiv:2512.07331v2 Announce Type: replace
Abstract: Vision Transformers (ViTs) lack the hierarchical inductive biases inherent to Convolutional Neural Networks (CNNs), theoretically allowing them to maintain high-dimensional representations throughout all layers. However, recent observations suggest ViTs often spontaneously manifest a "U-shaped" entropy profile-compressing information in middle layers before expanding it for the final classification. In this work, we demonstrate that this "Inductive Bottleneck" is not an architectural artifact, but a data-dependent adaptation. By analyzing the layer-wise Effective Encoding Dimension (EED) of DINO-trained ViTs across datasets of varying compositional complexity (UC Merced, Tiny ImageNet, and CIFAR-100), we show that the depth of the bottleneck correlates strongly with the semantic abstraction required by the task. We find that while texture-heavy datasets preserve high-rank representations throughout, object-centric datasets drive the network to dampen high-frequency information in middle layers, effectively "learning" a bottleneck to isolate semantic features.
arXiv:2607.14301v1 Announce Type: new
Abstract: As generative AI (GenAI) becomes increasingly embedded in undergraduate academic writing, how students rely on these tools, rather than simply whether they use them, has become a central question for learning, academic integrity, and educational equity. Existing measures of reliance were developed inductively, focused on discrete problem-solving tasks, and validated mainly with homogeneous samples. This study developed and validated the GenAI Reliance Types Scale (GenAI-RTS), a 20-item instrument measuring four theoretically derived types of GenAI reliance: Strategic, Instrumental, Dependent, and Dialogic. Validation followed the multisource framework of the Standards for Educational and Psychological Testing, drawing on a survey of 382 undergraduates at a U.S. Minority-Serving Institution and interviews with 14 purposively sampled students. Confirmatory factor analyses of six competing models supported a five-factor structure in which Strategic Reliance comprises two facets, Deliberate Use and Critical Evaluation, alongside Instrumental, Dependent, and Dialogic factors (CFI = .92, RMSEA = .08; DWLS CFI = .98, RMSEA = .07). Subscale reliability was acceptable to good (omega = .75-.88), and scalar measurement invariance held across gender, first-generation status, and STEM/non-STEM majors, to our knowledge the first such evidence for a GenAI reliance instrument. Rasch analysis indicated that a five-point response format would improve category functioning. Strategic reliance was positively associated with AI literacy, and the reliance types differentiated students across multiple writing process and outcome variables. The GenAI-RTS offers researchers and educators a theoretically grounded, psychometrically validated instrument for identifying undergraduate reliance profiles and supporting research, assessment, and AI literacy intervention.
arXiv:2607.15199v1 Announce Type: new
Abstract: We present a theory of a flow of salt solutions in long axisymmetric channels induced by concentration and pressure drops between their ends. The consideration is restricted to thin, compared to the local radius, electrostatic diffuse layers, but remains valid even when the concentration drop is quite large. We show that the magnitude of the diffusio-osmotic fluid flow rate $Q_{DO}$ in the cylinder is the same as in the slit of equal to its diameter thickness, but channels of variable cross-sections could either retard or enhance it, depending on their geometry. The application of the pressure drop $\Delta p \neq 0$ results in an extra contribution $Q_{P}$ to the total flow rate of fluid $Q$, but does not affect $Q_{DO}$. We calculate the curves $\Delta p (Q)$ for several axisymmetric channels and conclude that they are nearly linear, with the sensitive to the shape slopes. This leads to the possibility of introducing a simple, but rather accurate, cylinder approximation, where the radius of the imaginary cylinder is related to a hydrodynamic resistivity of the real channel and can be easily determined, if its geometry is known. We also derive an equation relating the ionic flux with the total flow rate of fluid and demonstrate that both the sign and magnitude of ionic flux could be tuned by using the appropriate channel shape. Our analysis provides a framework for interpreting experimental and numerical data, as well as may guide the design of micro- and nanofluidic devices.
arXiv:2607.14753v1 Announce Type: new
Abstract: Recent advances in text-to-speech and voice cloning make high-quality spoofing inexpensive and scalable, threatening voice authentication systems, especially automatic speaker verification (ASV). Existing defenses mainly address this threat through binary countermeasures (CMs) for deepfake detection or spoofing-aware speaker verification (SASV), where current systems are dominated by modular ASV-CM fusion and cascaded pipelines. Although large audio language models (LALMs) have shown promise on related audio tasks, including CM and ASV, their use for SASV remains unexplored, despite their capacity to produce natural-language rationales for auditing and robustness beyond discriminative predictions. This work systematically evaluates LALMs for SASV against conventional pipelines under zero-shot prompting, supervised adaptation, reasoning-oriented training, and reinforcement-learning-based optimization. Our results show that pretrained LALMs are near chance in the zero-shot setting, confirming that they are not natively suited to SASV, but that task-specific adaptation closes this gap. We further find that competitive SASV performance can be achieved through several distinct routes. These findings position LALMs as a promising and auditable foundation for unified SASV, while clarifying where conventional cascade systems still lead.
arXiv:2607.15005v1 Announce Type: new
Abstract: As advanced packaging technology evolves, increasing interconnect density in redistribution layers (RDLs) makes routability critical to package floorplanning. Meanwhile, power integrity requirements often reserve fan-in regions for the power delivery network (PDN), forcing signal nets through fan-out regions and complicating routability estimation. Existing uniform grid-based congestion models cannot accurately characterize fan-out congestion, while previous pin assignment methods struggle to evaluate net crossings. We propose a differentiable routability-driven floorplanning and pin assignment algorithm for advanced packaging with fan-out routing. First, a differentiable wirelength minimization method directly models discrete chip orientations and back-propagates wirelength gradients to chip locations and orientations. It reduces wirelength under fixed pin selection while avoiding the bias of continuous-angle modeling. Second, a crossing-aware pin assignment method incorporates net-crossing cost into a multi-strategy DPSO algorithm and uses GPU-parallel cost evaluation to reduce wirelength efficiently. Finally, a differentiable routability maximization method constructs a congestion model tailored to fan-out routing and establishes a back-propagation path from congestion information to chip locations, thereby guiding routability optimization. Experimental results show that our method achieves 100% routability on all benchmarks. For cases successfully routed by the baselines, it reduces wirelength by up to approximately 23% compared with a leading floorplanning method equipped with our pin assignment flow.
arXiv:2607.14757v1 Announce Type: new
Abstract: This paper determines the structural and proof-theoretic consequences of the forcing condition in relational quantum modal logic, under which every modal transition available at a world is also available at every world compatible with it. We prove that modal successor sets are constant on compatibility components, so boxed truth sets belong to the Boolean algebra of unions of these components. The relation holding exactly between worlds in the same component assigns to each stable proposition its greatest lower and least upper approximations by unions of components, and in hard superselection models these are exactly the approximations by central propositions. We adopt local validity for sequents with multiple conclusions to give a semantics for modal excluded middle on frames with several components. A connectedization obtained by adding one point then shows that component frames, equivalence frames satisfying the forcing condition, and connected compatibility frames with universal modal accessibility have the same logic for sequents with one conclusion. Finally, maximal consistent pairs yield a canonical model, and the calculus obtained by adding T, 4, and B is proved sound and complete for the three frame classes. These results characterize the logical scope of the forcing condition and establish a complete proof theory for component modalities.
arXiv:2607.14142v1 Announce Type: new
Abstract: We present an outlier-robust Bayesian approach for automated kinetic profile fitting in magnetically confined plasmas with the modified tanh (mtanh) parametrisation and demonstrate its implementation on KSTAR. The method addresses two systematic obstacles: anomalous diagnostic channels can bias least-squares fits, and multimodality of the mtanh cost surface can trap deterministic optimisers in secondary minima. The deployed workflow uses a good-and-bad Gaussian mixture likelihood based on the Box--Tiao formulation as the default outlier-robust likelihood for fitted diagnostic channels, with posterior outlier probabilities retained as channel-level quality indicators. The posterior is sampled with an affine-invariant ensemble MCMC sampler initialised near the result of deterministic maximum a posteriori (MAP)-seeking optimisation, reducing sensitivity to secondary minima on the multimodal mtanh surface. A batch automation layer retrieves diagnostic data from MDSplus and fits arbitrary time slices in parallel for the quantities \(n_e\), \(T_e\), \(T_i\), and \(v_T\) for which the relevant diagnostics are available. Results are written in formats suitable for MDSplus upload and downstream analysis. Representative KSTAR H-mode cases show that the mixture likelihood downweights contaminated measurements while preserving plausible pedestal profiles. The workflow provides a practical basis for future large-scale kinetic profile production for kinetic-EFIT, TRANSP, FASTRAN, and data-driven analysis workflows.
arXiv:2607.15058v1 Announce Type: new
Abstract: CAD-to-image alignment aims to estimate an object's 9D pose (rotation, translation, and anisotropic scale) from a single RGB image, enabling applications in robotics and augmented reality. Recent zero-shot methods use visual foundation models to match image regions to CAD models, yet typically their correspondences are appearance-driven and degrade under occlusion or sim-to-real domain shift. To address these limitations, we introduce SUFLECA (Scaling Up Feature LEarning for CAD Alignment), a weakly-supervised framework for zero-shot CAD alignment with two key contributions. First, SUFLECA scales up geometry-grounded feature learning from pretrained visual representations through Normalized Object Coordinates (NOCs) supervision on 674K images spanning 12 real and synthetic datasets, learning compact geometry-aware features that generalize across domains. Second, we propose a geometrically consistent matching algorithm that establishes reliable one-to-one CAD-to-image correspondences. Together, these contributions enable accurate, sub-second alignment per object instance without iterative pose refinement. On ScanNet25k, SUFLECA achieves 33.4%/42.3% category/instance accuracy, outperforming, with a smaller computational footprint, the strongest zero-shot baseline by 10.3/12.2 percentage points and, for the first time on this benchmark, even surpassing fully supervised methods. Code is available at: https://github.com/snt-arg/SUFLECA
arXiv:2607.14167v1 Announce Type: new
Abstract: LLM agents often retry after external validation rejects a candidate, but the interface between validation and the next model call remains underspecified. We introduce VeriHarness, a code-controlled agent loop in which models generate candidates while external validators control acceptance, budgets, and traces. We use it to compare raw diagnostics with feedback that identifies the failure location, observed value, and admissible alternatives. Across 50 paired TextWorld games under a four-call cap, feedback containing all three fields raises terminal success from 14/50 to 36/50 for Qwen2.5-Coder-14B (+44 percentage points) and from 8/50 to 29/50 for Llama-3.1-8B (+42 points). Ablations locate most of the gain in the admissible alternatives: feedback containing only the location and observed value remains near the raw diagnostic baseline. Presenting the complete repair information in prose instead of a keyed JSON record yields nearly the same success, providing no evidence that JSON syntax itself improves repair. The ordering persists across the tested call budgets and one sampled-decoding setting.
arXiv:2607.15122v1 Announce Type: cross
Abstract: Hysterons provide a minimal description of memory in driven matter: bistable elements with distinct switching thresholds whose interactions generate hysteresis, avalanches, and return point memory or its violation. Experimental realizations have so far been dominated by solid state mechanical systems, where bistability is usually encoded structurally through buckling, snap through, or geometric incompatibility. Here we realize hysteron physics through a hydrodynamic route. A single elastic fiber anchored in a microfluidic channel becomes bistable through nonlinear elastohydrodynamic feedback: viscous loading deforms the fiber, deformation reshapes hydraulic resistance, and flow redistribution modifies the loading. This feedback produces a fluidic hysteron whose onset is organized by a cusp catastrophe in geometric control parameters. A parallel bypass channel acts as a geometric load line that reshapes, and can even eliminate, bistability while simultaneously mediating long ranged hydraulic interactions between fibers. In arrays, varying a single geometric parameter drives a transition from a non interacting Preisach regime with return point memory to an interacting regime with avalanche like switching and return point memory violation. These results establish a passive hydrodynamic route to hysteron networks, in which memory emerges from flow structure feedback and global hydraulic constraints rather than solid state multistability or external control.
arXiv:2607.14841v1 Announce Type: cross
Abstract: We present a set of ultra-compact spectrographs fabricated using two-photon polymerisation, each under 1 mm in length and printed directly onto an optical fiber. This fully integrated, alignment-free architecture enables extremely low-mass instrumentation with broad potential across astrophotonics, space systems, drone-based sensing, and quantum optics. Two transmissive configurations are presented; they combine beam-expansion optics, dispersive elements, and stray-light suppression within monolithic printed structures optimised for J- and H-band operation. Coarse gratings provide resolving powers of approximately R = 30. Experimental characterisation confirms the feasibility of fiber-coupled, fully printed spectrographs and demonstrates their promise as deployable components for next-generation integrated photonic instrumentation.
arXiv:2607.14862v1 Announce Type: cross
Abstract: In this paper, we propose a novel tamed stochastic gradient Hamiltonian Monte Carlo (tSGHMC) algorithm for sampling and stochastic optimization problems with superlinearly growing stochastic gradients. Under a certain continuity in average condition and a strong convexity condition, we establish a non-asymptotic error bound in Wasserstein-2 distance for tSGHMC with the rate of convergence equal to $1/4$. Then, we derive an upper estimate for the associated expected excess risk, which provides a theoretical guarantee for the performance of tSGHMC. To illustrate the effectiveness of the proposed algorithm, we apply tSGHMC to practical examples, including a newsvendor problem and a Conditional Value-at-Risk minimization problem, using synthetic and real-world datasets. Numerical results support our theoretical findings. Furthermore, we compare tSGHMC with its first-order counterpart, namely, the tamed unadjusted stochastic Langevin algorithm. Simulation results demonstrate that tSGHMC achieves lower root mean square error and expected excess risk across a range of tasks.
arXiv:2508.16307v2 Announce Type: replace
Abstract: Metamorphic testing is a widely used methodology that examines an expected relation between pairs of executions to automatically find bugs, such as correctness bugs. We found that code coverage cannot accurately measure the extent to which code is validated and mutation testing is computationally expensive for evaluating metamorphic testing methods. In this work, we propose Metamorphic Coverage (MC), a coverage metric that examines the distinct code executed by pairs of test inputs within metamorphic testing. Our intuition is that, typically, a bug can be observed if the corresponding code is executed when executing either test input but not the other one, so covering more differential code covered by pairs of test inputs might be more likely to expose bugs. While most metamorphic testing methods have been based on this general intuition, our work defines and systematically evaluates MC on five widely used metamorphic testing methods for testing database engines, compilers, and constraint solvers. The code measured by MC overlaps with the bug-fix locations of 50 of 64 bugs found by metamorphic testing methods, and MC has a stronger positive correlation with bug numbers than line coverage. MC is 4x more sensitive than line coverage in distinguishing testing methods' effectiveness, and the average value of MC is 6x smaller than line coverage while still capturing the part of the program that is being tested. MC required 359x less time than mutation testing. Based on a case study for an automated database system testing approach, we demonstrate that when used for feedback guidance, MC significantly outperforms code coverage, by finding 41\% more bugs. Consequently, this work might have broad applications for assessing metamorphic testing methods and improving test-case generation.
arXiv:2607.15218v1 Announce Type: new
Abstract: Large language models (LLMs) increasingly serve as high-level planners for embodied agents, where linguistically benign instructions can become unsafe once grounded in the physical world. We study whether this physically grounded danger is the same safety problem as ordinary text-level content danger. Through hidden-state direction analysis and random-split null tests, we show that content danger (CD) and physical danger (PD) form separable signals in LLM representations across Qwen2.5-3B/7B/14B/32B, Phi-3.5 and SmolLM2. Building on the CD/PD separability, we propose PRISM, a single-layer L2-regularized logistic probe over full hidden states. PRISM achieves 86.2--87.7\% accuracy on SafeAgentBench with 11.7--13.7\% FPR, while same-scale LLM judges over-block safe tasks at 24.7--39.0\% FPR. We further introduce PhysicalSafetyBench-1K (PSB-1K), a contrastive benchmark of 1{,}000 physical-risk pairs without direct harm keywords, to test whether methods detect physically grounded danger rather than explicit unsafe wording. On PSB-1K, PRISM reaches 99.6\% accuracy and 0.7\% FPR, whereas a Qwen2.5-3B judge rejects 67.8\% of safe tasks. PRISM also replicates on SafeText and EARBench, supporting hidden-state probing as a representation-level method for physical safety beyond text moderation.
arXiv:2607.14770v1 Announce Type: new
Abstract: Knowledge graph question generation (KGQG) aims to generate natural-language questions from structured graph evidence. Existing KGQG benchmarks, however, are mostly built on static knowledge graphs and do not encode the temporal scopes of graph facts. As a result, they cannot evaluate whether generated questions faithfully preserve temporal validity, event ordering, and answer-determining temporal constraints. In this paper, we study temporal knowledge graph question generation (TKGQG), where a generated question must be faithful to both the support subgraph and the temporal constraints required to identify the target answer. We propose ChronoQG, the first temporally expressive and hop-bounded benchmark construction framework for TKGQG. ChronoQG integrates a comprehensive temporal-constraint taxonomy, topology-temporal subgraph sampling, and trace-grounded question generation to construct temporally faithful questions. The framework produces four benchmark datasets from heterogeneous temporal knowledge graphs, totaling 16,011 verified questions. We evaluate representative LLM-based KGQG methods and prompting baselines across diverse TKGQG settings, including temporal-constraint counts, topological templates, and temporal-constraint types. The results show that existing methods struggle to preserve temporal constraints, especially under multi-constraint settings and harder temporal-constraint types. These findings reveal a clear gap between static KGQG and TKGQG, and establish ChronoQG as a challenging testbed for temporally faithful question generation.
arXiv:2607.15239v1 Announce Type: new
Abstract: Coenen \textit{et al.}\ (\textit{J. Fluid Mech.}, vol.~921, 2021, p.~R2) developed a reduced-order model of peristaltic pumping in non-axisymmetric annular conduits with rigid walls, in the context of periarterial space (PAS) flows. \textit{In vivo} studies show that the PAS's outer wall undergoes significant displacement due to flow within and that the penetrating PASs form a porous pathway. To account for these biomechanical aspects, we revisit the problem of flow in an eccentric annular conduit and incorporate porous drag and two-way-coupled fluid--structure interaction between the compliant outer wall and the cerebrospinal fluid flow within. A Darcy--Brinkman term in the axial momentum equation accounts for drag due to the porous medium. We account for changes in hydraulic resistance due to peristalsis and compliant-wall displacements perturbatively, thereby reducing the problem to a single nonlinear partial differential equation for the axial pressure. This reduced-order model allows us to build a mechanistic understanding of flow through a porous penetrating PAS and enables parametric studies. For small-amplitude peristaltic waves, analytical solutions are possible.
arXiv:2607.15241v1 Announce Type: new
Abstract: Healthcare multimodal AI must combine visual and textual evidence while remaining reliable and interpretable. Using MediaEval Medico 2025 as a retrospective GI endoscopy case study, we analyze design choices across nine documented systems for question answering and explanation quality. Parameter-efficient adaptation of pretrained backbones provides strong challenge performance, but answer-level gains do not consistently translate into faithful and complete clinical reasoning. Methods enforcing structured reasoning and explicit grounding show more reliable behavior across heterogeneous question types, although the evidence is correlational rather than ablation-based. These results motivate evaluation beyond lexical overlap, standardized evidence-linked explanations, leakage-aware data governance, and lightweight robustness and calibration checks. The findings support trustworthy multimodal healthcare AI based on data fusion, explainability, and resilient evaluation.
arXiv:2603.15727v3 Announce Type: replace
Abstract: Autonomous LLM-based agents increasingly operate as long-running processes forming densely interconnected multi-agent ecosystems, whose security properties remain largely unexplored. Systems such as OpenClaw, an open-source platform with over 40{,}000 active instances, persistent configurations, tool-execution privileges, and cross-platform messaging, are deployed at scale, yet the security of such agent ecosystems remains largely unexplored. This work presents AgentWorm, the first self-replicating worm attack against a production-scale agent framework, achieving a fully autonomous infection cycle initiated by a single message: the worm first hijacks the victim's core configuration to establish persistent presence across session restarts, then executes an arbitrary payload upon each reboot, and finally propagates itself to every newly encountered peer without further attacker intervention. The attack is evaluated on a controlled testbed across five distinct LLM backends, three infection vectors, and three payload types. Results show a 63\% aggregate attack success rate, sustained multi-hop propagation, and stark divergences in model security postures, highlighting that while execution-level filtering effectively mitigates dormant payloads, skill supply chains remain universally vulnerable. Defenses are evaluated at three layers (prompt-level mitigations sourced from real community practice, the framework's built-in security controls, and an ecosystem-wide measurement of public configurations), revealing that the critical controls capable of breaking the infection loop are not enabled in any of the observed deployments. A cross-framework transferability experiment on Hermes Agent confirms that the underlying vulnerabilities are properties of the autonomous agent design pattern, not artifacts of a single implementation.
Step-Tagging: Toward controlling the generation of Language Reasoning Models through step monitoring
arXiv:2512.14332v2 Announce Type: replace
Abstract: The field of Language Reasoning Models (LRMs) has been very active over the past few years with advances in training and inference techniques enabling LRMs to reason longer, and more accurately. However, a growing body of studies show that LRMs are still inefficient, over-generating verification and reflection steps. To address this challenge, we introduce the Step-Tagging framework, a lightweight sentence-classifier enabling real-time annotation of the type of reasoning steps that an LRM is generating. To monitor reasoning behaviors, we introduced ReasonType: a novel taxonomy of reasoning steps. Building on this framework, we demonstrated that online monitoring of the count of specific steps can produce effective interpretable early stopping criteria of LRM inferences. We evaluate the Step-tagging framework on three open-source reasoning models across standard benchmark datasets: MATH500, GSM8K, AIME and non-mathematical tasks (GPQA and MMLU-Pro). We achieve 20 to 50% token reduction while maintaining comparable accuracy to standard generation, with largest gains observed on more computation-heavy tasks. This work offers a novel way to increase control over the generation of LRMs, and a new tool to study behaviors of LRMs.
arXiv:2512.21306v2 Announce Type: replace
Abstract: This work systematically investigates the performance of FORCE--$\alpha$ numerical fluxes within an arbitrary high order semidiscrete finite volume (FV) framework for hyperbolic partial differential equations (PDEs). Such numerical fluxes have been recently introduced by Toro, Saggiorato, Tokareva, and Hidalgo (Journal of Computational Physics, 416, 2020), and constitute a family of centred fluxes obtained from a suitable modification of First--Order Centred (FORCE) numerical fluxes. In contrast with upwind fluxes, such as Rusanov, Harten--Lax--van Leer (HLL) or the exact Riemann solver (RS) numerical flux, centred ones do not consider in any way the structure of the Riemann problem at cell interfaces. Adopting centred numerical fluxes leads to a high level of flexibility of the resulting numerical schemes, for example in the context of complicated hyperbolic systems, for which RSs may be impossible to construct or computationally expensive.
The baseline framework adopted in this investigation is a FV semidiscrete approach with Weighted Essentially Non--Oscillatory (WENO) spatial reconstruction and Deferred Correction (DeC) time discretization, and results are reported up to order 7. Previous investigations involving the same framework have established that increasing the order of accuracy tends to decrease the differences in the results obtained through different numerical fluxes. The goal of this paper is to show that the employment of FORCE--$\alpha$ numerical fluxes within such a framework is a competitive alternative to the adoption of more classical upwind fluxes. The hyperbolic system considered for this investigation is the ideal Euler equations in one and two space dimensions.
Craig-Lyndon Interpolation for the Logic of Here and There with a Variation of Mints' Sequent System
arXiv:2601.04080v4 Announce Type: replace
Abstract: We present a variation of Maehara's method to construct Craig-Lyndon interpolants for the three-valued propositional logic of here and there (HT), also known as G\"odel's $G_3$, a superintuitionistic logic of importance in logic programming. Our method adapts a recent interpolation technique that operates on classically encoded logic programs to a variation of Mints' sequent system for HT. The approach is characterized by two stages: First, a preliminary interpolant is constructed, a formula that is an interpolant in some sense but not yet the desired HT formula. In the second stage, an actual HT interpolant is obtained from this preliminary interpolant. With the classical encoding, the preliminary interpolant is a classical Craig-Lyndon interpolant for classical encodings of the two input HT formulas. In the presented adaptation, the sequent system operates directly on HT formulas, and the preliminary interpolant is in a nonclassical logic that generalizes HT by an additional logic operator.
arXiv:2607.14894v1 Announce Type: cross
Abstract: Plug-and-play proximal gradient descent (PnP-PGD) enables flexible image reconstruction by using denoisers as implicit priors. In practice, these denoisers are often deployed outside their training domains. Existing analyses establish convergence under structural assumptions on the deployed denoiser, such as requiring it to be a proximal map or a contraction. However, they do not measure how domain mismatch affects convergence of PnP-PGD. We define this effect as \emph{proximal mismatch}: the discrepancy between a deployed denoiser $\widehat{\mathsf D}$ and a target-domain reference map $\mathsf D_\star=\operatorname{prox}_{R_\star}$ associated with the underlying regularizer $R_\star$. Under this mismatch, each denoising update becomes an inexact proximal step for the target objective. We further derive a stationarity bound that decays at a rate of $\mathcal{O}(1/K)$, with an additive term proportional to the average squared proximal mismatch. This result motivates adaptation via proximal matching rather than MSE-based adaptation alone. We study this approach with two established denoiser families: learned proximal networks and gradient-step denoisers. Experiments on Gaussian deblurring and super-resolution under substantial domain shift show that proximal matching adaptation improves reconstruction quality significantly over MSE-based adaptation, yielding the largest numerical gains in the few-shot regime.