Papers

Large audio-language models (LALMs) exploit multimodal evidence, yet task-irrelevant audio can alter text-reasoning decisions when listening is unnecessary. Aggregate Accuracy can hide this paired drift because audio-induced repairs and damages may cancel. Paired drift analysis and targeted interventions identify architecture-specific, intervention-sensitive late audio pathways as actionable control points. We introduce ICAP-Gate, which applies mechanism-guided, task-conditioned control to each model's pathway. Across four LALMs, two reasoning benchmarks, and environmental-sound and natural-speech interference, ICAP-Gate has lower point estimates for Influence Rate and Answer Flip than ungated inference in all 16 full-split model–condition evaluations. Fixed suppression degrades automatic speech recognition (ASR) across all four models, whereas ICAP-Gate matches ungated ASR performance by preserving the pathway for explicit audio-demand instructions. ICAP-Gate has lower paired-drift point estimates than mitigation prompting in all four evaluated settings and provides competitive stabilization relative to eight-sample Self-Consistency while using one generation per query; in controlled ARC measurements, Self-Consistency incurs $7.0$–$9.2\times$ ungated latency. These results establish selective modality influence control as a design principle for robust multimodal reasoning.

Quantum compilers play a critical role in transforming high-level quantum programs into optimized, hardware-compatible circuits. However, verifying the correctness of compiler passes remains challenging, as determining the expected output of large, deeply entangled quantum circuits is computationally intractable. This challenge is further amplified when compiler passes modify already complex circuit structures, making manual validation of transformed circuits impractical. In this work, we perform a systematic analysis of unit tests for quantum compiler passes in four quantum programming frameworks (PennyLane, Qiskit, Cirq, and pytket). Our findings indicate validation is dominated by program-content and program-metric assertions, and test circuits are generally small and shallow. Motivated by these observations, we introduce a testing methodology for automated validation of quantum compiler passes based on retromorphic testing and principles from the Hadamard test. This methodology analyzes a compiler pass, test circuit, and expected pass behavior to verify semantic preservation and intended structural modifications. We implement our methods in a framework, RetroQ, and apply it to compiler passes in PennyLane and Qiskit. Experimental evaluation reproduced several existing bugs as well as uncovered previously undetected defects, such as flawed symbolic parameter handling, incorrect commutation logic, failure to recognize self-adjointness of gates, and runtime crashes. These findings highlight the need for compiler-pass-specific testing methodologies to improve the reliability of the evolving quantum software stack.

Dexterous grasping is the foundational primitive in embodied AI, demanding massive data to train robust models. As real-world data collection is expensive, simulation has become the mainstream paradigm. Yet, while cluttered scenes best reflect real-world applications, learning to grasp within them is bottlenecked by a critical scarcity of large-scale data. To resolve this, we curate high-quality 3D objects and supporting bases, proposing a scalable seed-and-filter strategy that bypasses sluggish scene-level optimization. This yields an unprecedented benchmark comprising over 2.6 million scenes and 0.4B scene-specific grasp ground truths, featuring diverse realistic layouts paired with rich semantic and geometric observations. Furthermore, we introduce the OmniDex model to overcome the grasp multimodality and last-millimeter precision errors plaguing current generative models. By coupling Soft Winner-Takes-All learning with human-inspired physical constraints during training, and utilizing physics-driven ranking, our approach achieves robust dexterous grasping without the latency of post-optimization. Experimental results show that OmniDex model achieves state-of-the-art performance and strong generalization across diverse scenes, views, and unseen objects.

We establish local existence and uniqueness of strong solutions to the compressible Navier-Stokes/Cahn-Hilliard system in a bounded smooth domain in $\mathbb R^3$, allowing the initial density to vanish. The initial data satisfy suitable regularity and compatibility conditions for the momentum and phase equations. The main difficulty arises from the simultaneous degeneracy of the momentum equation and the density-weighted phase equations, which complicates the construction of compatible positive-density approximations and uniform initial estimates. We overcome this difficulty by introducing a time-independent residual into the chemical relation. The residual preserves the phase compatibility exactly, keeps the initial phase field unchanged, and vanishes in $H^1$ as the approximation parameter tends to zero. For each positive-density approximation, we construct a local strong solution using a semi-discrete Galerkin scheme that discretizes only the phase variables and retains the full transport and momentum equations. We then derive a priori estimates independent of the density lower bound, which yield a common lifespan and allow us to pass to the vacuum limit. Uniqueness is proved using a modified difference energy that accounts for the coupling between the density and the chemical potential. To our knowledge, this is the first local well-posedness result for the three-dimensional compressible Navier-Stokes/Cahn-Hilliard system with initial vacuum.

We investigate the complementarity between collider searches and gravitational-wave (GW) observations in the classically scale-invariant general $U(1)_X$ extension of the Standard Model. The $U(1)_X$ scale is generated radiatively through the Coleman-Weinberg mechanism, while the electroweak scale is induced through the Higgs portal, without explicit mass terms in the scalar potential. Three right-handed neutrinos are introduced to ensure anomaly cancellation and acquire Majorana masses upon $U(1)_X$ symmetry breaking, which also generates the mass of the neutral gauge boson $Z'$. In the strongly supercooled regime, a tachyonic $U(1)_X$ phase transition can produce a stochastic GW background. Assuming three heavy Majorana neutrinos, we study their effects on the phase-transition dynamics and reheating, and estimate the resulting GW signals. We identify the regions of the $U(1)_X$ gauge coupling and $Z'$ mass accessible to future GW observatories, including LISA and DECIGO, and compare them with existing LEP and LHC constraints. We find that GWs from tachyonic phase transitions can probe regions with small gauge couplings and heavy $Z'$ bosons that are difficult to access through collider searches, demonstrating the complementarity of these probes.

Spin–orbit torque MRAM (SOT-MRAM) is a leading candidate for next-generation nonvolatile memory, offering high speed, endurance, and architectural compatibility. However, in conventional SOT switching, the magnetization remains near the in-plane region at pulse termination, making the final state highly sensitive to post-pulse relaxation dynamics and prone to back-switching. To overcome this, we propose a field-free scheme in which the transverse SOT drives large-angle precessional excitation while the perpendicular spin-transfer torque (STT) biases the trajectory toward the reversed $-z$ state. Micromagnetic simulations reveal a nonlinear switching boundary in the $J_{\mathrm{STT}}$–$J_{\mathrm{SOT}}$ parameter space, originating from the distinct dynamical roles of the two torques: SOT primarily governs the excitation and crossing of the dynamical separatrix, whereas STT controls the terminal trajectory and final-state selection. An analytical macrospin model, based on the stability analysis of the current-induced equilibrium, reproduces the critical-boundary trends as functions of current density, Gilbert damping $\alpha$, and uniaxial anisotropy $K_\mathrm{u}$, and distinguishes dynamic anti-damping and static instability branches. Systematic analyses of pulse duration, damping, anisotropy, and the STT–SOT balance further demonstrate that reliable ultrafast switching requires not only sufficient excitation to cross the separatrix before pulse termination, but also precise control of the pulse-end magnetization state to minimize post-pulse relaxation. These results establish that the pulse-end state, rather than the instantaneous torque amplitude, is the decisive factor governing switching speed and reliability in coupled STT–SOT systems.

Navigation integrity and estimator credibility use error and uncertainty information to address different questions. This perspective distinguishes actual positioning error from reported statistical characterizations, formulates integrity as control of hazardous unwarned-use risk under the requirement of a specified operation, and describes credibility as the justification of an uncertainty or risk claim by its evidence usage and modeling process. Three hypothetical scenarios show how deteriorated accuracy, a contradicted uncertainty report, and a hazardous event can lead to different credibility and integrity judgments. The resulting assessment asks what happened, why the reported claim should be believed, and whether the information is sufficient for the intended operation.

We show that classical cylindrical contact homology over $\mathbb{Z}$, equipped with its decomposition by free homotopy classes, is not invariant under changes of hypertight contact form. Specifically, we construct two nondegenerate hypertight perturbations of the standard Seifert contact form on $\Sigma(2,4,5)$ whose cylindrical contact homologies in the regular-fiber class are respectively torsion-free and contain $\mathbb{Z}/5$-torsion.

Persistent memory can improve personalization in LLM agents but can also induce sycophancy and cross-domain leakage. We distinguish two governance decisions: admission, which determines what recalled information enters the working context, and presentation, which determines how admitted information is expressed. We implement two inference-time designs without retraining: factor-compiled admission (FC), which assesses whole memory entries, and permission-semantic admission (PS), which decomposes entries into typed units; both translate adjudicated attributes into eligibility decisions via deterministic policies. We evaluate on a four-backbone development suite and an external benchmark with four tasks of 300 samples each. Relative to verbatim injection, FC and PS reduce pooled judge-assessed failure rates on the external benchmark by 6.7 and 8.8 percentage points (p = 2.7e-7 and 4.1e-12), and development-set cross-domain leakage falls by up to 29.5 percentage points. A query-conditioned gating baseline shows no significant change in objective-fact failure or pooled failure. Under matched admission budgets, PS outperforms random and relevance-based selection on external objective-fact judgment after Holm correction. Holding presentation fixed, tightening admission cuts cross-domain failure by a further 17.5 percentage points (p = 1.6e-4); in contrast, no comparison between two renderings of identical adjudicated outputs survives multiple-comparison correction. Both designs increase personalization failures, and PS misses the preregistered improvement and personalization-preservation criteria. These results support evaluating admission and presentation separately: selection quality provides task-specific safety gains, while preserving beneficial memory use remains unresolved.

Single-view 3D reconstructors often produce plausible meshes that disagree with the input view, especially near depth discontinuities and self-occlusions. We present a lightweight, plug-and-play RGBD-to-3D refinement that improves any RGB-to-3D reconstructor without retraining. Given a depth map, we correct the visible surface by bipartite matching to back-projected depth points, mirror these corrections onto the occluded side across a detected symmetry plane, and propagate them with a smoothness solver. Every stage is closed-form, making the method orders of magnitude faster than optimization-heavy test-time refinement. On GSO and OmniObject3D with five backbones, it yields consistent gains, also with monocular pseudo-depth, benefits more from symmetry on symmetric objects, and compares favorably with prior refinement in accuracy and runtime. It further improves an RGB-D-to-mesh reconstructor and transfers to real captures with noisy sensor depth.

E-values provide a flexible framework for statistical inference, but the universal threshold $1/\alpha$ can be conservative when the null distribution has additional structure. We study how structural restrictions limit the worst-case concentration underlying Markov's inequality, using a nondecreasing density model to constrain tail allocation and an $L$-Lipschitz density model to control local concentration. The nondecreasing model gives the minimax rejection threshold, and the Lipschitz condition yields a sharper closed-form threshold with an $O(L^{-1/2})$ relative improvement over $1/\alpha$. We further show that this threshold is asymptotically sharp as $\alpha\to0$, in the sense that no uniformly valid threshold can improve on it by a fixed positive $L$-dependent amount. We then incorporate the Lipschitz calibration into multiple testing procedures with FDR control and use the same structural information to construct calibrated conditional e-values for anytime-valid sequential inference. Numerical experiments illustrate the resulting gains in discoveries and the trade-offs of the sequential procedures.

Single-cell trajectory inference maps transcriptomic measurements onto developmental continua, yet configurations that fit the data equally well can assign conflicting cell fates. FateMultiplicity is a label-free framework that constructs a statistically admissible model set, or Rashomon set, without lineage labels, by evaluating model discrepancy on cross-fitted held-out genes under non-inferiority testing calibrated against random-seed variation. Multiplicity is large and depends more on the diversity of the model space than its size: twelve configurations of a second algorithm expose 20.0% of cells where twenty-four of the first expose 3.8%. Whether the per-cell certified fate margin FM yields more reliable assignments than the fitted model already provides is then tested, and it does not. On simulation ground truth, on the same cells, FM discriminates misassignment at AUC 0.682, against 0.965 for the baseline configuration's own decision margin (p = 0.003) and 0.854 for a seed-dispersion baseline. Informativeness is governed by the breadth of the admitted set, not its cardinality: at cardinality four, seed refits give 0.933 and hyperparameter-perturbed sets 0.701. Relaxing the infimum to a q-quantile recovers discrimination but converges toward the single model's own confidence; the supremum reaches 0.973 because theta*'s membership bounds it from below, while the infimum is unanchored. Multiplicity in trajectory inference is worth measuring and reporting, but per-cell certification over a label-free Rashomon set is not a route to more reliable fate calls. Two constructions survive: a margin-erosion ratio separates real from spurious branch points in simulation (AUC 0.890, untested on real data), and against clonally observed fate, uncertified cells disagree with their clone's outcome 16.4 percentage points more often than certified cells (p < 0.001).

Advances in image generation have made visual authenticity increasingly difficult to assess. Although image forensics now examines both generation artifacts and higher-level visual inconsistencies, a plausible image can still contradict real-world facts or rules. We introduce WorldFact-Bench to evaluate image-world consistency from a single image, without a predefined claim or verification target. The benchmark contains 1,274 source-aligned real-fake pairs across four verification regimes and ten semantic domains. Each pair introduces a specific, evidence-supported factual conflict while seeking to preserve non-target content and visual plausibility. Images are evaluated independently, and pair accuracy requires both members of a pair to be classified correctly. We further propose PERSIST-Agent, which organizes iterative verification around a persistent state linking candidate facts, visual observations, evidence, and verification statuses. This state guides subsequent inspection and retrieval while retaining unresolved candidates. With backbone weights fixed, harness self-optimization refines the agent's prompts and execution rules through validation feedback. Experiments reveal strong label biases in several detectors and uneven gains from retrieval. On the evaluated 8B backbones, PERSIST-Agent improves pair accuracy over both direct judgment and retrieval-augmented baselines, while ablations support the role of persistent verification state. These findings highlight the value of state-guided verification and the remaining gap between visual plausibility and factual correctness.

Following retrieved evidence does not guarantee factual correctness: misleading evidence can induce a model to replace an answer it previously gave correctly. Standard accuracy measures obscure this behavior by combining answer replacement with preexisting errors. We introduce RAG-Stress, a controlled diagnostic protocol for examining the limits of evidence reliance in retrieval-augmented generation. The protocol holds the question and reference answer fixed, edits one assertion to support a designated incorrect answer, and crosses two source priority policies with three positions of the answer span within the evidence text. We measure misleading rate (MR) on each model's subset of questions answered correctly without retrieval, alongside clean accuracy on the full evaluation set. We evaluate fifteen systems spanning API models, open models, and search agents trained with reinforcement learning on TriviaQA-RC, HotpotQA, and SearchQA, with additional English and Chinese MedQA evaluations. Instructions that prioritize documents consistently produce higher MR than those permitting reliance on prior knowledge. Averaged over models and positions, the gap ranges from 10.9 to 13.5 percentage points across the three QA datasets. Mean MR follows End $>$ Beginning $>$ Middle under both policies, although individual models do not uniformly follow this ordering. A separate paired audit of 500 questions and two checkpoints supports increased harmful override without establishing a corresponding improvement in beneficial correction. These findings distinguish evidence adherence from factual reliability and motivate evaluating whether retrieved evidence preserves, replaces, or corrects a model's answers.

Data driving is an important technique for modelling realistic magnetic field and plasma evolution in regions of the solar atmosphere where direct observations are either absent or insufficient. This technique uses observational data to drive the evolution of the simulation, while the modelled solar atmosphere evolves self-consistently. A common approach for data-driven simulations is B-driving, which uses the observed magnetic field at the surface of the Sun as a boundary condition. However, only prescribing B at the boundary does not generally determine its numerical evolution when the update depends on the reconstructed interface states or fluxes. To address this, E-driving provides an alternative approach by instead driving the simulation using the electric field on the boundary, which directly controls the magnetic field update. We systematically quantify the performance of B- and E-driving in a controlled set of numerical experiments. We test both methods in the Athena++ solver using (a) an analytical flux-emergence model (Y. Fan & S. Gibson 2003) and (b) a "ground-truth" reference simulation (S. Toriumi & S. Takasao 2017; S. Toriumi et al. 2020) of active region emergence. For E-driving, the electric field is incorporated into Athena++'s constrained-transport (CT) update, preserving the discrete solenoidal constraint. We assess the reproduction of (i) boundary magnetic flux, (ii) relative magnetic helicity, (iii) total and free magnetic energy, and (iv) magnetic morphology. Across these diagnostics, E-driving gives better overall reproduction of the reference evolution than B-driving. For the configurations considered in an Eulerian CT scheme, E-driving is more accurate than the commonly used ghost-cell B-driving strategies and simpler than constructing an update-consistent B-driving boundary condition.

Generative models have rapidly pushed content creation be-yond 2D imagery toward dynamic 3D and 4D scene synthesis. Yet pro-ducing realistic and temporally stable 4D content from a single image is still difficult because one view provides limited structural cues and weak motion evidence. We introduce MATE4D, a framework that converts one input image into editable dynamic 4D content. Our method constructs a spatio-temporal multi-view image matrix with text-guided background manipulation, delivering coherent supervision over viewpoint, appear-ance, and motion. These synthesized observations are used to optimize 3D Gaussian primitives, which are then animated through a lightweight deformation module to form a 4D representation. The resulting scenes preserve geometry more faithfully, maintain smoother temporal behavior, and keep background edits more consistent, reducing context ambiguity and motion artifacts. Experiments on Objaverse-XL and Diffusion4D show that MATE4D outperforms strong baselines in visual quality, effi-ciency, and controllability, supporting practical AR/VR content creation.

Kucher asked whether $\ell_{\infty}(E)$ is a Grothendieck space for every super-reflexive Banach space $E$. We answer this question affirmatively by proving the stronger result that $\ell_{\infty}(E)$ has Pełczyński's property $(V)$. Combined with the natural dual representation of $\ell_{\infty}(E)$, this yields that every non-weakly compact operator on $\ell_{\infty}(E)$ fixes a copy of $\ell_{\infty}$. The main ingredient is a general property-$(V)$ theorem for function modules over zero-dimensional compact Hausdorff spaces with stalks admitting a common modulus of uniform convexity. The proof is based on Gierz's integral representation of functionals and the weak compactness of the associated canonical measures. In the case of $\ell_{\infty}(E)$, the relevant stalks are ultrapowers of $E$, and super-reflexivity provides the uniform geometric control required by the theorem. We also derive several consequences for operators on $\ell_{\infty}(E)$, including weak compactness into separable spaces, reflexivity of separable quotients, and compactness of operators into Schur spaces.

AI coding agents are moving from local code assistance into pull-based workflows, where generated contributions must be reviewed, explained, and maintained within existing project norms. Although recent work has begun to characterize AI-authored pull requests (AIPRs), less is known about how reviewers govern their entry into review, how AI authorship reshapes credibility and fairness, and what intake mechanisms protect review sustainability. We report a mixed-method questionnaire survey of 239 practitioners from 31 countries with code-review experience and varying exposure to AIPRs. In the scenarios and self-reports elicited by the survey, AI authorship was not a categorical rejection signal. Instead, respondents described review effort as conditional on whether an AIPR arrived as an accountable contribution, with bounded scope, project-grounded rationale, validation beyond Continuous Integration (CI), contributor responsiveness, and identifiable post-merge ownership. This conditional logic extended to newcomer AIPRs, where respondents emphasized visible participation in the current review process over profile-level reputation alone. Qualitative responses further described shifts in mentoring, scrutiny, deferral, and routing when human stewardship was difficult to observe. We conceptualize missing rationale, validation, and ownership as reviewability debt, the work reviewers must absorb when generated code lacks sufficient human grounding. These findings reframe AIPR governance around making human judgment observable before generated contributions consume scarce reviewer attention.

We report an accurate optical frequency synthesizer, which can generate single-frequency laser light with high frequency stability and accuracy at desired frequencies over a wide optical region. The frequency of the output signal is divided from an 171Yb optical lattice clock via an accurate optical frequency divider based on an optical frequency comb. Therefore, the output of the optical frequency synthesizer inherits the frequency accuracy from the Yb optical clock. The frequency uncertainty of the 171Yb optical lattice clock is evaluated to be 6.8 mHz, corresponding to a fractional frequency uncertainty of 1.3E-17, mainly limited by the blackbody radiation shift and the lattice-induced light shift.

This paper develops a general framework for dynamic mechanism and information design with AI agents, encompassing optimal contracting and integrating evolving hidden information and hidden action with partial verification, technological authorization, and multi-agent interaction. It nests conventional dynamic screening, hidden-action contracting, and conditional dynamic information design, while allowing reporting, action, and information-use decisions to interact through a common continuation mechanism. We establish a dynamic revelation principle and a recursive characterization of implementability in which truthful reporting and post-report obedience are jointly determined through continuation incentives. In quasilinear optimal-contracting environments, dynamic envelope and integral-monotonicity methods yield a verification-adjusted dynamic virtual-surplus representation and identify three distinct margins: dynamic information rents, verification rents, and the implementation cost of hidden action. These margins generate a branched hierarchy separating conventional dynamic screening, hidden-action contracting, and their joint problem. We also study the joint design of monitoring, information use, and incentives: more informative monitoring capacity expands the designer's opportunity set, but full disclosure need not be optimal, and information use can itself alter hidden-action incentives. With multiple agents, correlated information, peer discipline, coupled incentives, joint feasibility, and continuation opportunities create additional strategic interactions. Although motivated by AI-agent environments, the framework applies more broadly to dynamic agency settings with similar economic features.

Multimodal remote sensing image registration is a crucial prerequisite for the collaborative processing and downstream application of remote sensing data, such as image fusion, change detection, and target recognition. However, significant variations in radiometry, geometry, scale, viewpoint, and time often exist between multimodal images. These differences, driven by varying sensor geometries, physical radiation mechanisms, imaging platforms, and environmental disturbances, pose severe challenges to achieving high-precision, robust registration. This paper systematically reviews the progress of mainstream multimodal remote sensing image registration methods. Based on their registration pipelines, existing approaches are categorized into three main types: region-based, feature-based, and deep learning-based methods. We detail the core principles, representative algorithms, advantages, and limitations of each category. Additionally, we summarize publicly available multimodal image datasets in the remote sensing domain, analyzing their specific characteristics and applicable scenarios. Finally, we highlight current bottlenecks in high-precision registration research and outline future development trends. This review aims to provide a comprehensive reference and valuable insights for researchers in related fields.

Modern data-driven decision-making methods, such as imitation learning (IL) and reinforcement learning (RL), have achieved great success in solving many complex tasks. However, these methods often suffer from serious control instability and robustness issues when applied in real-world applications such as robotics and autonomous driving, posing notable challenges for their practical deployment. We argue that this instability issue stems largely from their limitations in solely supervising and optimizing zeroth-order actions (i.e., the action labels), failing to account for higher-order action dynamics and temporal consistency. In this paper, we show that simultaneously supervising both zeroth- and first-order actions can dramatically enhance policies' performance and control robustness. To achieve this, we introduce a novel and elegant loss scheme supported by formal theoretical guarantees that can equip any off-the-shelf policy model (e.g., deterministic, stochastic, or flow policies) with the capability for higher-order action supervision, without requiring any structural modifications. Moreover, our proposed method can serve as a lightweight plug-and-play module that seamlessly integrates with a broad spectrum of existing offline RL frameworks. Extensive evaluations on OGBench and D4RL demonstrate that our approach yields substantial performance and robustness improvements across a wide range of continuous control environments. Notably, our method can also enhance policies' out-of-distribution (OOD) generalization capability in the challenging low-data regime, making it an ideal tool in tackling many real-world control problems.

While recent video generation models synthesize highly realistic visuals, they lack a genuine understanding of intrinsic real-world logic. Existing methods attempt to understand the world by internalizing diverse world knowledge, yet constrained by computational overhead or dimensionality alignment, their learning processes inevitably compress features, causing a severe loss of structural information. To address this, we propose IntactWorld, a Joint World Modeling Architecture utilizing uncompressed Intact Features. Since data naturally reside on a low-dimensional manifold within a high-dimensional space, predicting the flow velocity $v$ within this uncompressed high-dimensional space induces a severe manifold gap. To successfully eliminate this optimization bottleneck, our framework instead predicts the clean feature $x_0$ at intermediate layers. Furthermore, to mitigate the computational overhead of incorporating complete world knowledge, we introduce a Full-to-Compact Training Paradigm. By replacing raw full features with highly refined CLS tokens, this paradigm enables efficient single-branch guidance, reducing spatial memory consumption by 11.4% and cutting inference latency by 43.8%. Extensive evaluations demonstrate the effectiveness of IntactWorld, outperforming established baselines by 2.46 points on the VBench 2.0 benchmark.

The likelihood input to a Bayesian analysis almost never exactly represents how the data were generated, calling into question the validity of posterior inferences. We review a variational interpretation of the Bayesian posterior as an alternative justification for its use under model misspecification, and consider the resulting implications on uncertainty quantification and parameter estimation. We then introduce a range of techniques that seek to obtain generalised Bayesian inferences that account for model misspecification

Integrated Sensing and Communication (ISAC) is envisioned to endow future 6G systems with seamless sensing capabilities. To support efficient sensing with minimum communication overhead, sparse pilots embedded within communication frames have emerged as a promising solution. Along this line of research, existing studies have achieved engaging results in maximizing the unambiguous sensing region. However, jointly maximizing the sensing region and sensing accuracy remains challenging due to the lack of a unified performance metric and an effective pilot design framework. This paper jointly optimizes the unambiguous sensing region and sensing accuracy for estimating delay-Doppler (DD) parameters in Orthogonal Frequency Division Multiplexing (OFDM)-ISAC systems, where sensing mutual information (SMI) is adopted as a unified performance metric to characterize the overall sensing capability. Specifically, the joint optimization is formulated as an SMI maximization problem by systematically resolving sensing ambiguity and optimizing sensing accuracy. In particular, based on the generalized Bezout identity, we derive a 2D (timefrequency) coprime condition, which, as far as the authors know, is the first necessary and sufficient condition to achieve the unique estimation of DD parameters in the literature. Under this unambiguous condition, we further propose an Adam-Guided Iterative Refinement (AGIR) algorithm to optimize the sensing accuracy. Numerical results demonstrate the advantage of the proposed framework over existing designs, owing to the freedom offered by the 2D coprime condition in optimizing the sensing accuracy.

Long-form video understanding often involves multiple questions about different aspects of the same recording. Yet existing video agents typically process each question through an isolated tool-use trajectory. This repeatedly restarts video exploration and memory construction, missing opportunities to acquire evidence jointly and progressively build a shared understanding that supports the complete question set. We introduce VAMR (Video Agent for Multi-Question Reasoning), which coordinates all questions about a video through one shared tool-use trajectory. At each round, a persistent policy model can invoke tools for one or more unresolved questions and submit answers for questions with sufficient evidence. Question-conditioned visual perception retrieves fine-grained clues for several questions in one call, while layered multi-question memory integrates reusable context into a shared video story and preserves separate evidence for individual questions. After supervised fine-tuning initializes this interaction protocol, we propose question-horizon policy optimization (\qhpo) to optimize shared trajectories in which questions progress and finish at different rounds. Specifically, a question-level critic estimates the value of each active question, while round alignment maps each question advantage to the rounds that directly serve it before the aligned advantages are aggregated to optimize the shared actor. Across LVBench, Video-Holmes, and LongVideoBench, VAMR achieves the highest accuracy overall and the fewest reasoning rounds among iterative methods. On LVBench, it reaches 62.1% accuracy, exceeding VideoARM by 4.3 points while reducing reasoning rounds and processed frames by 85.9% and 61.4%.

Long-term time series forecasting (LTSF) models predominantly employ patch-based encoders terminated by a flatten readout head that maps the entire encoded historical memory to all future steps through a single shared projection. This implicit coupling of future positions obscures position-specific historical-to-future alignment and amplifies sensitivity to corrupted inputs and extreme supervision noise. We present SACQ, a plug-in structured prediction head that replaces flatten readout while keeping the encoder unchanged. SACQ adopts a two-stage decoding pipeline: it first establishes a coarse patch-grid forecast scaffold, then refines each future position through cross-attention over historical memory and merges the attention-derived correction with the coarse scaffold via a learned per-patch gate. To stabilize optimization under long horizons and noisy labels, we further propose a batch-adaptive scaled log-cosh loss that automatically calibrates robustness to the current residual scale, suppressing outlier gradients while preserving MSE-like sensitivity for typical errors. SACQ attains top-tier test MSE/MAE across PatchTST, DLinear, and patch-Mamba backbones with only modest incremental overhead in parameters and latency. Under inference-time input corruption and training-set label-noise stress tests, SACQ substantially outperforms flatten readouts, with ablation studies validating each architectural component.

Agent skills are SKILL.md instructions and scripts that AI coding agents such as Claude Code and Codex run with the permissions of their user. Developers share skills by copying them between repositories, which makes them a software supply chain without a registry, versions or provenance. The origin of a copied skill, the reach of a security fix and the repositories that warrant review are therefore unknown. Studies that record which repositories hold a skill at a single point in time cannot reveal who copied it from whom. We contribute the first dated copy network of agent skills, built from the git history of every SKILL.md in GitSkills and covering 2,193,119 skill adoptions across GitHub, together with an interactive viewer. A few repositories are the source of almost all copies, and GitHub stars do not identify them. Skill copies almost never change with their source, and a fix at the source therefore rarely reaches them. We fit a model of which repositories others copy from and use it to rank repositories for audit. Reviewing the 100 repositories it ranks highest prevents 14.9% of later adoptions of high-risk skills, against 0.5% for the 100 most starred, which gives security engineers a short list to check before a skill spreads. Platforms should therefore distribute versioned references rather than copies. Project Website: https://fahdseddik.github.io/Skill-Constellations/

Robotic manipulation often contains repeated motions whose local observations look similar at different phases. When these phases require different actions, a policy that relies mainly on the current observation may repeat completed motions or switch phases at the wrong time. To address this phase ambiguity, we present the Pseudo-Memory Temporal Re-encoding Module (PMTRM), a lightweight plug-in module with only 7.61M parameters that encodes a bounded history of executed states and actions into a latent sequence for existing policies. To help distinguish phases, a temporal heterogeneity objective penalizes positive similarity between distant positions in this sequence, while anchor and reconstruction losses preserve information needed for action prediction. The reconstruction decoder is used only during training, leaving the temporal re-encoder to supply history to the policy at inference. We train the module progressively on synthetic sequences and robot data, then jointly with the policy, using temporal masking to accommodate partial histories. This integration retains the original action head and action space and adds auxiliary losses to the original policy loss. Experiments with multiple policy backbones in simulation and on a real robot show improved task success on tasks with phase ambiguity, with little additional computation.

Vertical-domain few-shot classification remains challenging for small language models, as limited supervision makes it difficult to acquire domain-specific decision knowledge. On-Policy Distillation (OPD) can improve teacher-guided adaptation by supervising student-generated rollouts, while GRPO-based reinforcement learning can further refine downstream predictions. However, existing KD-to-RL pipelines typically rely on globally fixed transition schedules, ignoring that different samples may require different amounts of teacher-guided acquisition before reward-driven refinement. We propose PIVOT (Perplexity-Informed Transition Optimization), a dynamic transition framework that routes samples between OPD and GRPO according to teacher-evaluated sequence perplexity. PIVOT moves low-perplexity samples to GRPO for reward-driven refinement while keeping high-perplexity samples under OPD for continued domain knowledge acquisition. Experiments on Banking77 and HWU64 show that PIVOT consistently outperforms continued OPD and globally synchronized OPD$\rightarrow$GRPO baselines under the same number of post-warm-up student optimization steps, achieving stronger downstream performance and more stable training dynamics.