Papers

Here,we report an ST plasma in which Thomson scattering measures a central electron temperature exceeding 8.6 keV (100 million degrees kelvin) on the EXL-50U spherical torus device.The high-temperature phase is obtained at low density by sequential operation of two 50 GHz ECRH systems: ordinary-mode (O-mode) ECRH3 followed by extraordinary-mode (X-mode) ECRH4. In contrast to neutral-beam-heated high-ion-temperature ST scenarios, the discharge is sustained by ECRH only.

Large language model (LLM)-based AI Scientists advance scientific discovery, yet they suffer from inefficiencies due to noisy observations. Existing principle-evolving agents act on every anomaly, leaving its source unresolved: a new mechanism or an evaluation artifact, so invalid revisions bias evolution. To address this, we formulate revision as evidence admission: anomalies steer search immediately, but only independently verified evidence may rewrite persistent principles, and contradicted principles are retracted. We realize this in TRIAGE-PIEVO , which pairs an anomaly-responsive search portfolio with a staged verification gate that alone controls what enters principle state. Across four scientific tasks with three backbones, TRIAGE-PIEVO (1) attains the highest average solution quality with +7.5%~11.8% over the strongest baseline per backbone, matching the strongest baseline's full-budget quality with one-third of the evaluation budget on Qwen3-235B; and additional stress tests and cross-harness deployment demonstrate that (2) in constructed stress tests, it admits none of the corrupted, hidden-source, or exploit evidence that ungated variants admit, while missing no planted mechanism, and (3) plugged unchanged into three external research harnesses, it cuts false evidence admissions by over 60% and raises verified scores by 31%~43%.

Vision-and-language navigation (VLN) agents are typically evaluated by whether they reach a target and how efficiently they follow an instructed route, yet these metrics provide little information about collision behavior. Moreover, continuous VLN evaluation commonly uses a sliding regime, where an agent can slide along obstacles after a collision, potentially masking the effect of unsafe actions on navigation performance. We introduce VLN-Safe, a benchmark for analyzing collision safety under two complementary evaluation regimes: sliding, reflecting common VLN evaluation practice, and blocking, where collision-inducing actions prevent forward motion and more directly expose their consequences. Across six published agents on R2R-CE, 38.5-56.8% of successful episodes contain contact under sliding, while switching to blocking decreases success rate by 11.7-36.1 percentage points. We further analyze signals for anticipating imminent collisions and find that depth geometry and recent movement history are more informative than policy-derived signals, while safe alternatives near the policy's preferred action are rarely available. Motivated by these findings, we introduce a training-free collision guard that vetoes risky forward actions, executes an escape maneuver, and returns control to the original policy. Under blocking, the guard reduces collision steps by 62-98% and improves success rate by 6.9-29.3 percentage points across five recurrent, topological, and vision-language-action agents. Under sliding, it leaves success rate unchanged while raising the share of collision-free successes by 4.1-12.4 points. These results reveal a substantial gap between task success and collision-safe navigation and show that lightweight collision-aware intervention can improve both safety and navigation performance.

In this paper we study pairwise compatibility graphs whose witness tree is required to be an unweighted caterpillar. This is a natural restriction of PCGs, in the same line of research as the study of star witnesses, which led to the class of double-threshold graphs. In the caterpillar case, we show that this restriction is equivalent to a one-dimensional difference-labelling problem: a graph admits such a representation if and only if its vertices can be assigned integer labels so that adjacency is determined exactly by whether the absolute difference of two labels belongs to a given interval. Using this characterization, we derive structural properties of the class, including bounded twin-width, and study its relation with familiar graph families. We show that the class contains cycles and two-dimensional grids, while it excludes natural obstructions such as the subdivided star $S_{2,2,2,2,2}$, the cube graph $Q_3$, the wheel $W_6$, and the complement of the $3$-sunlet. We also consider restricted representations, including centipede witnesses and representations with a given interval and we give a polynomial-time recognition algorithm when the interval is fixed.

Sign Language Recognition (SLR) presents complex challenges due to variable signing speeds and spatial intricacies. Specifically, Indian Sign Language (ISL) requires precise modeling of its inherently bimanual nature and is bottlenecked by the scarcity of large-vocabulary benchmarks and the inability of standard models to handle irregular signing velocities. To address this, a structurally-aware hybrid framework for efficient, gloss ISL recognition is proposed. To minimize computational redundancy, an entropy-driven SVD sampler that dynamically extracts informative keyframes is used. Because this subsampling creates irregular physical time gaps, the proposed architecture employs a Continuous Dual Positional Encoding paired with a distance-aware Elastic Transformer to preserve natural signing rhythm. Spatially, a Kinematically-Conditioned Adaptive GCN explicitly models bimanual geometry using unified bone-joint features. To validate our framework, a 120-class, 9,453-video native ISL dataset is created. Augmented by Physical Node Dropout and ArcFace optimization to mitigate subject-specific memorization, our framework achieves a competitive 89.8% accuracy, establishing a strong baseline for complex sign language modeling.

We introduce HumanPhys, a human-centric benchmark for real-world physical perception, and a training corpus emitted by the same pipeline. Each question targets a physically grounded quantity related to people, including depth, inter-person distance, gaze angle, displacement, velocity, heading-step trajectory, blink count, and contact-deformation strength. The rendered probes additionally test illumination inference from body and face cues and camera ego-motion, both with and without a person in the frame. All labels are derived programmatically from sensor calibration, source metadata, or renderer state, rather than written by annotators or produced by estimators on the benchmark frames. HumanPhys systematically probes six dimensions of human-centric physical perception: spatial state, kinematics, blink events, deformation, illumination, and camera pose. To the best of our knowledge, no existing MLLM benchmark-and-training corpus jointly supervises this set of human-centered physical quantities under direct metadata- or renderer-derived labels. The six dimensions appear only partially dissociable across the tested MLLMs. We benchmark a range of open- and closed-source MLLMs on HumanPhys and find that all six capability dimensions remain challenging for current models. We hope HumanPhys will support future research on multimodal systems that perceive and reason about people in the physical world.

We study upper tail large deviations for the first passage time in the frog model on $\mathbb Z^d$, $d\ge 2$. Let $\mathrm{T}(0,x)$ denote the first time at which the infection, started from the origin, reaches $x$, and let $\mu$ be the associated time constant. We prove that, for any Euclidean unit vector $v\in\mathbb R^d$ and $\xi>0$, \[ \lim_{n\to\infty} -\frac{1}{\mathfrak{s}_d(n)} \log \mathbb{P}\bigl(\mathrm{T}(0,\lfloor nv\rfloor) \ge (\mu(v)+\xi)n\bigr) =\mathfrak{r}_d\xi, \] where $\mathfrak s_2(n)=n/\log n$ and $\mathfrak r_2=\pi$ for $d=2$; $\mathfrak s_d(n)=n$ and $\mathfrak r_d$ is the escape probability of simple random walk for $d\ge3$. Moreover, on the upper tail large deviation event, we show that a certain slowdown mechanism occurs near the endpoint. To rule out other mechanisms, we analyze the growth of the infection near the starting point using a modified exploration process, and investigate its propagation through the intermediate region using the renormalization and covering arguments.

Moving operands to a processor can cost more than the arithmetic performed on them, and for small binary similarity and search workloads data movement can dominate. We present HADES, a programmable processing-near-memory engine built around one shared 32x32-bit 2R1W register-file macro that holds both data and program. Twelve 16-bit opcodes, three working registers and a 32-slot program express ten search workloads in 8 to 19 slots, and SCANP and SCANSEL shorten dual-row and selected scans. A post-layout 2x2 Tiny Tapeout implementation closes timing at 40 MHz with 26,968 square microns of standard cells beside a 15,744 square microns macro. All HADES results come from RTL or gate-level simulation of this implementation, and HADES energy is estimated. Against ESP8266 and ESP32 microcontrollers running plain C, and with an ideal host, HADES answers a resident query 10x to 32x and 2.2x to 11x faster, respectively. End to end, it wins at the evaluated batch sizes of B = 16 and B = 256 against both boards and loses the single-query (B = 1) runs to the ESP32. An Arduino Nano runs a restricted subset. Kernel power stays within about 2% of a spin loop on all three boards. These results hold for plain-C baselines under the evaluated conditions. At this scale the control and instruction storage fit around a shared memory macro, while the byte-wide interface and the loading path remain significant costs.

We present CAST-ASD (Causal Attention over Speakers and Time), a causal joint model for active speaker detection (ASD). Recent ASD models improve performance by jointly processing all faces in a scene, but operate offline, with the full clip available before any frame is labelled. Low-latency causal detectors instead process each face independently. CAST-ASD combines joint face reasoning with fully causal inference. It applies scene-level attention across faces within the current frame and causal temporal attention over each face's history, enabling information sharing across speakers with zero lookahead. Its parameter count does not grow with the number of faces. On MSDWild, where most negative face-frames contain a visible competing speaker, CAST-ASD reduces competing-speaker false alarms by 32% relative to the causal per-face detector it extends, with larger gains as scene occupancy increases. It outperforms all causal baselines on both validation splits while using only one-tenth the inter-speaker parameters of a three-block fixed-roster convolution. Code and models are available at https://github.com/prathamg007/CAST-ASD.

We show that gravity can remove the linear magnetic instability of a holographic plasma with an $SU(2)$ global symmetry. In the probe limit, charged vector bosons become tachyonic in their lowest Landau level above a critical field and condense into a vortex lattice. In Einstein-$SU(2)$ Yang-Mills theory in AdS$_5$ we include the backreaction of the field exactly, expanding only in the condensate around the D'Hoker-Kraus magnetic brane. The critical field diverges as the coupling $\alpha = \kappa^2/\hat g^2$ approaches $\alpha_\star = 2/3$, above which the normal phase has no linear charged instability at any field or temperature: we show this at zero temperature and establish it numerically at finite temperature. At zero temperature the field cancels from the mass of the lowest Landau level in the brane's AdS$_3 \times \mathbb{R}^2$ throat, and the instability is a violation of the throat's Breitenlohner-Freedman bound. The onset temperature vanishes at $\alpha_\star$ with an essential singularity of Berezinskii-Kosterlitz-Thouless type. For two-derivative Einstein-Yang-Mills duals in $d = 4$, 5 and 6 dimensions the normal phase has a charged instability if and only if $C_J/C_T < 8/(d^2(d+1))$, one tenth in $d = 4$. Above $\alpha_L = 0.372$ the gravitational attraction between vortices makes the uniform lattice unstable and the transition first order, so $B_c$ is only a spinodal; whether a condensed phase survives above $\alpha_\star$ is left open.

We developed a staff rostering system for complex employment and staffing requirements at a municipal water utility and deployed it in routine operations at the Water Purification Office of Suita City Waterworks Bureau, Japan, in May 2026. We estimated that preparing a monthly roster took approximately 2-3 h after deployment, compared with 12-18 h previously. The system represents roster states, effective staffing and office assignments using binary and auxiliary integer variables and optimises the entire planning horizon. When the normal model yields no feasible solution, a second stage attempts to generate a reference roster by penalising weekly day-off shortages while retaining the other hard constraints. Outputs identify shortages and meeting-date changes applied before solving. Computational experiments used operational inputs for 14 staff over 245 days. The baseline was solved to optimality in approximately 177 s; a modified meeting-day scenario produced a reference roster with five staff-days of shortage. Independent monthly solving without history produced 16 staff-week cases of excess days off across two month-boundary weeks. This case demonstrates the practical value of optimisation-based decision support for staff rostering in municipal water utilities by showing how mathematical optimisation can be integrated into roster preparation and checking workflows with reduced preparation time.

User-generated audio collections frequently contain textual metadata such as titles, tags, and descriptions. Effectively leveraging these complementary fields requires accounting for both field het erogeneity and varying semantic relevance relative to the acoustic content. We propose FAM-Gate, a two-stage framework for metadata-enriched audio classification. In the first stage, field-aware metadata-view attention dynamically aggregates multiple metadata views using audio-guided weighting, explicit field-identity embeddings, a shared identity-initialized projection, and mean-anchored regularized aggregation. In the second stage, a prior-constrained vector gate performs bounded, dimension-wise fusion of audio and metadata representations centered around an acoustic prior. Ex periments on BSD10K, BSD35K, and FSD50K show consistent improvements in both hierarchical single-label and multi-label classification. Compared with matched random-view scalar-fusion base lines, FAM-Gate improves the hierarchical F-score by 2.53 and 5.72 percentage points on BSD10K and BSD35K, respectively, while increasing macro mAP and macro F1 by 5.07 and 10.25 percentage points on FSD50K.

Training-free sparse attention is a practical route to efficient video diffusion transformer inference, yet existing methods typically fix the competition domain at each query row for every attention head. Prior work primarily addresses how to estimate token-level importance within this fixed domain; whether the domain itself should adapt to head heterogeneity remains under-specified. We introduce EG-Sparse, a training-free framework that routes low-entropy heads to a grouped cross-head pool while retaining per-row local selection for high-entropy heads, using only scores already materialized by the sparse pipeline. A controlled diagnostic fixes the ranking score and varies only the competition rule: grouped cross-head selection reduces low-entropy-head reconstruction error by 9.8%, whereas forcing high-entropy heads into an all-head global pool increases it by 7.9%. Across four configurations, EG-Sparse achieves higher video quality while reducing attention density and GPU denoising latency compared with strong training-free sparse-attention baselines. Mechanism ablations further show that selective domain adaptation outperforms uniform, random, and fixed domain-assignment policies. These results support competition-domain adaptation as a distinct design axis within centroid-guided training-free sparse attention.

Streaming video understanding requires answering questions from the observed video, with evidence drawn from recent observations or earlier events. Historical context can improve recall but may interfere with judgments based on recent frames. Its usefulness depends on both when it is introduced and which clips are selected, while relevance alone does not guarantee sufficient evidence for an answer. We propose SPARE, a framework that combines adaptive historical retrieval with visual encoder sharing. A lightweight gate predicts whether recent observations provide sufficient evidence to answer the question. When additional evidence is predicted to be necessary, a multimodal embedding model retrieves relevant historical clips to supplement the current context. LoRA instruction tuning trains the understanding model to use recent observations and retrieved history jointly and abstain when the evidence remains insufficient. To reduce redundant visual parameters, the understanding and retrieval branches share a visual encoder while retaining separate language backbones. Four linear adapters align the shared features with the retrieval model through distillation at the visual token and final retrieval embedding levels. With visual-only inputs, SPARE achieves 77.78% on the evaluated Real-time and Backward task groups of OVO-Bench and 73.31% on StreamingBench. Compared with Qwen3-VL-8B-Instruct evaluated at 2 fps, SPARE improves accuracy by 20.68 percentage points on OVO-Bench and 14.15 percentage points on StreamingBench.

Dividing each trajectory's score contribution by its realized length can change the direction of a reward-centered policy update. In a finite trajectory model, we derive an exact ascent threshold for a three-outcome softmax policy and exhibit a binary-reward example in which the expected Euclidean update decreases task success. Natural-gradient ascent holds for fully controllable outcome distributions with deterministic binary labels, but can fail under a one-parameter restriction. A third-order U-statistic estimates the first-order return change without bias from independent current-policy trajectories, one per request. In the example, a same-sample plug-in estimator instead has the opposite sign in expectation for batch sizes two through forty-nine. Exact policy-grid enumeration confirms the local sign reversal. In a branching process, length-normalized learning initially loses return, then recovers and exceeds the unnormalized comparator. The analysis concerns the expected update field at the current policy.

We consider stable and effective restarts of the SS–RR and SS–RRR methods for computing the eigenpairs of a large Hermitian matrix $A\in\mathbb{C}^{n\times n}$ with the $n_{ev}$ eigenvalues in a given interval. Taking the CJ–SS–RR and CJ–SS–RRR methods for instance, we prove that the commonly used restart scheme and its orthonormalized variant, named the natural restart schemes, are not numerically stable, thereby lowering attainable accuracy of approximate eigenpairs and missing some of the $n_{ev}$ desired eigenpairs. To cure these deficiencies, we propose a new restart scheme, called the combination restart, that takes the $n\times \ell$ starting matrix to be the product of the matrix, which consists of the selected $n_{ev}$ approximate eigenvectors, and an $n_{ev}\times \ell$ Gaussian random matrix. We prove the remarkable mathematical and numerical superiority of the combination restart over the natural restarts, then present a hybrid restart scheme that fully exploits the advantages of the natural restart and combination restart, and finally develop CJ–SS–RR and CJ–SS–RRR algorithms with the hybrid restart. Numerical experiments confirm that the new algorithms are numerically backward stable to the working precision and more efficient than the CJ–SS–RRR algorithm with the natural restart. The combination restart and hybrid restart schemes can be straightforwardly extended to the SS–RR and SS–RRR methods for general standard and generalized eigenvalue problems.

We propose a crossed electric spin generation effect in multilayer structures. This effect combines the advantages of in-plane and gate electric fields for manipulating spin, and unveils a nice quantum geometric structure in terms of the underlying hybrid Berry curvatures that are unique to quasi-2D systems and manifest inherent spin-layer-orbital coupling. Combining our theory with first-principles calculations, we predict sizable spin generation in 2D ferromagnetic Fe3GeTe2 and altermagnetic V2Te2O. Our theory opens a new Berry-curvature route for all-electric spin manipulation in 2D materials

We reduce the problem of establishing pointwise upper decay estimates in thin domains for viscosity subsolutions of fully nonlinear partial differential equations to solving a one-dimensional differential inequality. The resulting estimates make explicit the interaction between the transverse geometry of the domain and the ellipticity of the equation, and are shown to be sharp. Our framework allows the ellipticity to degenerate at infinity and accommodates nonhomogeneous equations with lower-order terms. As an application, we obtain decay estimates for $p$-subharmonic functions when $1<p\leq 2$

Precise reconstruction of particle decay chains is essential for a broad range of analyses in particle physics experiments, particularly those targeting flavour physics and spectroscopy. We present a standalone, modular and extensible decay tree reconstruction framework designed to handle complicated topologies, accounting for tightly constrained particle decays and long-distance trajectory extrapolations through regions of intense and inhomogeneous magnetic field, as well as limited momentum resolution. Implemented in C++, the framework is optimised for configurability and performance. Developed primarily within CERN's LHCb experiment, its focus is on decay chains with vertices located several metres from the LHC proton-proton interaction point, where the dipole magnetic field is maximal and non-uniform. Even so, it remains flexible: it can accommodate these special cases as well as the approximations commonly used in standard decay tree fitting tools, and the geometries of general-purpose collider detectors. Benchmark studies using simulated data demonstrate significant improvements in reconstruction efficiency, invariant mass resolution, and vertex position resolution with respect to existing tools, opening new opportunities for precision measurements and searches involving highly displaced decays.

Generative AI can be a powerful tool for supporting the design and implementation of quantum applications, whereas the burden of checking whether generated circuits realize the intended computation increases. In this study, we regard this burden of human understanding, checking, and judgment as “trust cost,” and propose a development methodology that integrates AI and formal verification to reduce it. In the proposed development methodology, the implementation is divided into three components: a target mathematical specification, a quantum circuit, and a correspondence theorem and its proof that prove the agreement between the two. This division separates the specifications and theorem statements checked by humans from the theorem proofs entrusted to mechanical checking. As a concrete development example of this proposal, we describe an implementation for the time evolution of a stochastic particle system and its results.

We present SPIRE, a spin- and photon-polarization-resolved particle-in-cell (PIC) code for self-consistent simulations of laser-plasma interactions in the strong-field quantum electrodynamics (QED) regime. The classical PIC framework combines a Yee-grid finite-difference time-domain Maxwell solver, a relativistic Boris particle pusher, shape-function particle-grid interpolation, Esirkepov charge-conserving current deposition, laser boundary injection, and a relativistic moving window in one, two, and three spatial dimensions. A vectorial Laguerre-Gaussian laser injector based on an angular-spectrum formulation supports tightly focused, nonparaxial optical orbital-angular-momentum modes. The strong-field QED module treats nonlinear Compton scattering and nonlinear Breit-Wheeler pair production as stochastic processes using spin- and polarization-resolved probabilities in the local constant-field approximation. Electron and positron spin evolution includes stochastic transitions, no-emission corrections, and Thomas-Bargmann-Michel-Telegdi precession with a field-dependent anomalous magnetic moment. Photon polarization is represented by the Stokes vector and evolves through nonlinear Compton scattering, pair production, vacuum birefringence, and vacuum dichroism. The implementation is verified with one-, two-, and three-dimensional benchmarks. Classical PIC results are cross-checked against SMILEI, and the spin- and polarization-resolved QED modules reproduce published results for polarized electron generation, vacuum birefringence and dichroism, and dense polarized positron production. These benchmarks verify the coupled classical PIC, strong-field QED, spin, and photon-polarization modules implemented in SPIRE.

4D Gaussian Splatting (4DGS) has emerged as a powerful dynamic scene representation, supporting real-time high quality rendering but suffers from prohibitive storage overhead. Existing methods typically attempt to solve this by treating compression as a per-scene optimization problem, limiting generalizability and applicability where low-latency compression is required. In this work, we leverage existing image and video codecs and propose a generalized post-training compression framework for 4DGS to address these issues. Specifically, we apply importance-based adaptive quantization to Gaussians according to their importance levels, effectively refining bitrate allocation while remaining fully compatible with existing codecs. To further improve compression efficiency, we utilize the statistical properties of 4DGS and propose motion- and importance-aware packing to reorganize attributes into codec-friendly 2D maps. Our approach can achieve over 40x compression rate in seconds while maintaining high quality, facilitating the practical deployment of 4DGS.

Self-improving agents use failure feedback to revise their code or instructions, but gains after selection do not establish target-specific repair or generalization. We measure directedness by comparing repair of a randomly selected target failure with repair of the parent's other failures. Across 145 eligible modifications in public Darwin Gödel Machine logs, we find no detectable target-specific advantage. Controlled instruction-revision experiments with small open models show that relevant feedback can improve target repair. In a second study with a pre-specified analysis plan, target feedback raised repair by 20 percentage points for Qwen3-4B relative to matched same-category feedback; the 7-point increase for Phi-4-mini was not significant. Target repair did not reliably imply held-out improvement. Held-out gains were concentrated in settings with identifiable, repairable failure mechanisms, including planted errors and natural tool-interface failures. Benefits of edit localization and regression screening varied across settings, while eight-step revision loops and GEPA ended at similar mean held-out accuracy. Evaluations should report target repair, transfer to related failures, and held-out net change separately.

Public advertising transparency relies on a minimum accountability condition: an account authorized by a publisher in ads$.$txt or app-ads$.$txt should appear in the applicable sellers$.$json registry. When an available registry omits the declared seller ID, the authorization remains visible and machine-readable, but its public verification path breaks. We call this failure the seller-identity gap. We measure the gap using 233.4 million normalized advertising-transparency records and an absent-only, coverage-aware evidence rule. A declaration enters the numerator only when the applicable registry is available and omits its seller ID; unavailable, malformed, confidential, and other non-evaluable states remain separate. The resulting row-level floor is 24.25% in a high-popularity, low-removal reference population and 30.71% among publishers associated with at least 100 copyright-removal requests, a 6.45 percentage-point increase. The gap then remains near one third at stricter cumulative thresholds. Standardizing the populations to common weights over 1,469 shared ad systems reduces the primary contrast to 4.65 points: about 72% remains within systems, while 28% forms a composition-associated residual. We connect these population estimates to recurring seller-account coordinates, byte-identical declaration configurations, and provenance-bearing case packets. We frame this declaration-centered investigative use of public advertising records as Ad Transparency Intelligence: it reconstructs declared supply-chain relationships, identifies where public verification stops, and prioritizes evidence for corroboration without equating correlation with attribution. Advertising transparency supports accountability and investigation only when visible authorizations remain publicly verifiable.

Artificial intelligence (AI) agents promise to extend what users can accomplish beyond their existing expertise, but they also raise the risk of producing plausible-looking work that users may struggle to evaluate. We ask how task-relevant expertise and trust in AI shape agent-assisted work. To isolate these factors, we conduct a think-aloud study in which 20 professionals with varying data-science expertise use the same agentic tool to complete the same data-science hiring assessment, used at a major U.S. company. We find that surface productivity can come at the expense of understanding: participants who delegated more completed more of the task, but were also more likely to produce invalid solutions. Experts more often resisted this tradeoff, sacrificing completion to decompose the task and scrutinize agent outputs. Our findings call for evaluating agentic assistance not only by how much it accomplishes, but by whether users remain able to evaluate the work they produce.

Streaming speech assistants can launch tools before users finish speaking, overlapping execution with incoming speech to reduce user waiting time. However, subsequent self-corrections can invalidate apparently stable arguments, causing stale execution whose irreversible effects persist even after the final call is corrected. We propose Repair-Aware Semantic Survival Gating (RASSG) to jointly control tool launch and cancellation during streaming speech. RASSG launches a tool when three consecutive speech recognition results arrive with identical nonempty arguments, then combines calibrated argument invalidity risk with execution costs to decide cancellation at the last feasible opportunity before effects become irreversible. Evaluation on ReCAST, FDB-v3, and Audio2Tool uses a controlled execution clock with fixed tool runtimes and cancellation parameters. On ReCAST's prospective-repair subset, RASSG reduces the stale execution rate from LocalAgreement-2's 50.00% to 3.15%, with a mean delay of 3.676s across successful trajectories. The stale reduction also holds against every early-execution baseline on both public speech datasets. Ablations with identical initial launches support active cancellation, showing the value of controlling ongoing calls before their effects become irreversible. Code and data are available at https://github.com/daijinhu/RASSG.

Conventional control approaches to optimization problems under uncertainty entail a practical compromise: deterministic controls offer robustness at the expense of pathwise accuracy while stochastic controls can adapt to individual realizations but suffer from the curse of dimensionality. To achieve both robustness and adaptivity, a neural operator-based approach is proposed in this paper for distributed elliptic optimal control problems with high-dimensional random inputs. A data-driven training algorithm is first established to learn the mapping from coefficient realizations directly to their optimal pathwise controls. To bypass the burden of generating training data pairs, a physics-informed training algorithm is then proposed and recast as weakly supervised learning using pseudo label targets produced by standard adjoint-based control updates. Numerical experiments with rough coefficients are conducted to validate the effectiveness and efficiency of our robust-pathwise controls, which outperforms the benchmark robust deterministic control on unseen coefficient realizations owing to the generalization capability and fast inference of trained models.

Can autoregressive video diffusion be accelerated further when only four denoising steps remain? Existing feature caching approaches struggle in this regime, as large feature changes between denoising timesteps undermine both direct reuse and lightweight predictors. We observe that features across video chunks remain strongly correlated at matching denoising timesteps, offering a complementary source of guidance. Building on this insight, we introduce TGuide, a framework that accelerates autoregressive video diffusion through a lightweight Anchor-Transport-Correct (ATC) predictor guided by historical sampling trajectories. Our ATC combines two complementary references: the current chunk's preceding-step features preserve its spatial structure, while the previous chunk's same-timestep features provide information at the target denoising stage. By integrating these references through spatial retrieval and adaptive correction, ATC improves velocity prediction accuracy despite the large gaps between denoising steps. To mitigate error accumulation across both denoising steps and video chunks, we train the predictor on its own autoregressive rollouts, allowing it to adapt to states shaped by its own earlier predictions. An optional confidence-aware scheduler further adjusts the quality-efficiency trade-off. Experiments on Self Forcing, Causal Forcing, and HY-WorldPlay show that TGuide preserves generation quality while achieving denoising speedups of up to $1.89\times$ over the original 4-step generators. On Self Forcing, TGuide generalizes from 81-frame training videos to 333-frame generation, approximately four times the training length, while exceeding the original model's VBench score.

We maximise expected reward per unit of marginal tail risk, a reward-to-CVaR ratio against a baseline. The standard Dinkelbach route re-solves a Rockafellar–Uryasev LP or QP with one auxiliary variable per tail scenario at each multiplier update. Our method eliminates the multiplier iteration and keeps scenario auxiliaries out of the master, while still evaluating tail risk to generate cuts. A Charnes–Cooper/Schaible homogenisation removes the reward/risk multiplier, and the remaining convex constraint is separated on demand: each cut coefficient is an Euler allocation vector, read off a ranking and an average. The result is a single convex program, finitely terminating at the global optimum, with size independent of S — an LP in $M{+}1$ variables, or a second-order cone program in $M{+}2$ for a concave quadratic numerator. The method runs 20–100 times faster than the exact Dinkelbach formulation and only needs single-digit cut counts.

Pose Graph Optimization (PGO) is a classical and widely applied problem in state estimation for robotics, playing a fundamental role in localization, mapping, and navigation. Yet, it is nonconvex in nature and may have many local minimizers, thus difficult to analyze. In this paper, we investigate the optimization landscape and establish a sufficient condition of global optimality for PGO. The ultimate goal of this paper is to theoretically confirm that the optimization landscape is benign and attractive around the global minimizer, if noise is smaller than an explicit and a priori bound. To reveal such an insight, this paper develops three main technical results. First, we identify a proper basin of attraction (BOA) containing global minimizer, and characterize its geometric property by showing that its radius is controlled by noise in an explicit way. Then, by exploring the first order optimality conditions of PGO, we prove that a local minimizer is globally optimal if a noise-controlled criterion is satisfied. Finally, by combining above two results, we prove that one and only one local minimizer exists in the BOA and this minimizer is globally optimal, when the noise is smaller than an explicit bound that can be verified a priori. Numerical simulations validate the proposed theory under various random cases, showing that the obtained conditions provide an effective criterion for certifying the global optimality of PGO.