arXiv:2607.13247v1 Announce Type: new
Abstract: Quadratic probing is one of the most widely used open-addressing hash-table schemes in practice, but after more than half a century, even its most basic performance guarantees remain poorly understood. In this paper, we revisit quadratic probing through the lens of a smoothed variant in which each key follows a random probe sequence where its $k$th probe is expected at offset $\Theta(k^2)$. This is simultaneously a toy model for better understanding regular quadratic probing and a natural hashing scheme in its own right. We analyse smoothed quadratic probing for both Robin Hood ordering and anti-Robin Hood ordering and reveal a surprising separation: At load factor $1-\varepsilon$, anti-Robin Hood achieves an expected query time of $\Theta(\log \varepsilon^{-1})$, which matches the conjectured expected average successful query time for regular quadratic probing, while Robin Hood falls short at $\Theta(\varepsilon^{-1/2})$. Our analysis generalises to degree-$d$ probing for any $d \ge 1$ with expected query time $O(\max(\log \varepsilon^{-1}, \varepsilon^{1-2/d}))$ for anti-Robin Hood and $\Theta(\varepsilon^{-1/d})$ for Robin Hood.
Finally, we go beyond smoothed analysis: using the probabilistic method, we show that for every $d \ge 2$, almost every random fixed-offset degree-$d$ probing sequence achieves expected query time $O(\log \varepsilon^{-1})$ under anti-Robin Hood ordering, simultaneously over all admissible table sizes and load factors. Thus, while quadratic probing itself remains elusive, we prove that essentially all quadratic-probing-like fixed-offset schemes achieve the ideal performance under the anti-Robin Hood ordering.
Science Journals
arXiv:2607.13249v1 Announce Type: new
Abstract: Safety critical autonomous systems often adapt by adjusting controller parameters while keeping the underlying architecture fixed. This strategy breaks down when shifts in sensing, resource availability, or component health invalidate the original structural assumptions. This work introduces a method in which system maintain an explicit, graph-based representation of their architecture and reason over it during operation. The system is modeled as a directed graph of physical, functional, and model based modules, with edges capturing information and control dependencies. Adaptation is posed as a joint optimization over architectural configurations and module parameters, subject to operational constraints using a Monitor-Analyze-Plan-Execute loop-based finite state machine. Performance degradation is isolated via residual decomposition and dependency weighted influence propagation, and candidate adaptations are filtered using a stability aware mechanism. The approach is demonstrated on a differential drive robot under sensor drift and actuator faults. A fixed architecture accumulates tracking errors of up to 24 m and 13 m, respectively, whereas architecture aware adaptation reduces error under 1.5 m in each case, by selecting fault appropriate configurations. These results show the value of reasoning over system structure, while preserving stability, rather than relying solely on parameter tuning.
arXiv:2607.13254v1 Announce Type: new
Abstract: Accurate and robust attitude estimation is a key challenge for autonomous vehicles, particularly in GNSS-denied conditions and during highly accelerated flight. In such conditions, Inertial Measurement Units (IMUs) alone are insufficient for reliable tilt estimation due to the ambiguity between gravitational and inertial accelerations. Although auxiliary velocity sensors such as GNSS, Pitot tubes, Doppler radar, or Visual Inertial Odometry are commonly used, they may be unavailable, intermittent, or costly. This paper introduces a barometer-aided attitude estimation architecture that exploits barometric altitude measurements to provide complementary information on the vehicle's vertical motion, thereby enhancing attitude estimation within nonlinear observers on SO(3). The contributions are twofold. First, we design a deterministic Riccati observer cascaded with a complementary filter, ensuring almost-global asymptotic stability (AGAS) under a uniform observability (UO) condition while preserving the geometric structure of the attitude dynamics. Second, we propose a nonlinear observer evolving on SO(3)xR2, which integrates IMU measurements as inputs and barometer and magnetometer measurements as outputs within a unified framework, guaranteeing local exponential stability (LES) under relaxed uniform observability conditions. The proposed approaches are validated using both simulated and real flight data. The results demonstrate that barometer-aided estimation provides a lightweight, reliable, and effective complementary sensing modality for attitude estimation in minimal-sensing configurations, offering a practical alternative when conventional velocity measurements are unavailable or degraded.
arXiv:2607.13262v1 Announce Type: new
Abstract: Accurate measurement of out-of-plane thermal conductivity remains challenging, particularly for multilayer thin-film materials, due to the reliance of existing techniques on indirect, surface-sensitive observables. Here, we introduce transient depth thermography, a non-contact spectroscopic method that directly reconstructs spatiotemporal temperature profiles using time-resolved thermal radiation. By leveraging the wavelength-dependent optical penetration depth, the technique encodes depth information into the emission spectrum, enabling direct access to internal heat transport dynamics. We demonstrate the method on fused silica and MgF2, achieving agreement with established techniques while attaining an uncertainty of 2-5%. In addition, we measure the temperature-dependent thermal conductivity of MgF2 over a broad temperature range. Applicable to both bulk and thin-film systems, transient depth thermography provides a general and accurate framework for thermal characterization of complex materials.
arXiv:2607.13274v1 Announce Type: new
Abstract: Reinforcement learning is increasingly being considered for controlling real-world systems, from fusion plasma and autonomous vehicles to drug discovery and drinking water treatment, where reliability is essential and tuning budgets are limited. Actor-critic algorithms share a set of design decisions, such as how the policy is updated, how it represents the distribution over actions, how its gradient is estimated, and how often it is updated relative to the value estimator. Using a control task derived from a real water treatment plant, we analyze over 33,000 experiments to determine how these components affect variability across runs and sensitivity to hyperparameters. Common defaults, such as Gaussian action distributions with pathwise gradient estimators, are among the least reliable configurations, whereas bounded distributions with adaptive update schedules remain robust across a wide range of settings. These findings offer empirical guidance to practitioners across scientific and engineering domains for understanding and making component-level decisions when adapting actor-critic methods to new real-world control settings.
Pezego-HITL: A policy-grounded large language model architecture for agricultural extension in Ghana
arXiv:2607.13934v1 Announce Type: new
Abstract: Large language models are increasingly deployed in agricultural decision-support settings, yet high-stakes crop protection in smallholder agriculture requires more than output-quality benchmarks. Over a two-year design and evaluation programme, we formalise policy-constrained large language model assessment as an adaptive compute allocation problem that jointly captures safety compliance, helpfulness, operational latency, and expert supervision workload. We introduce P-EVAL (Policy-grounded Expert-calibrated VALidation protocol), a unified evaluation framework for policy-grounded decision support, evaluating the architecture on a simulated field query database consisting of 1,240 cases. The protocol is instantiated on the Pezego advisory architecture (Pezego-HITL) and evaluated in Ghana. Following offline judge calibration against gold-standard human expert decisions ($\kappa = 0.77$), we evaluate the architectural performance under simulated query workloads. Under P-EVAL, our memory-routed architecture improves the Policy Alignment Rate (PAR) to 0.94 and the Agronomic Utility Rate (AUR) to 0.95, while reducing P95 latency by 55% (from 28.6s to 12.9s) through a 59.6% cache reuse ratio. We also demonstrate generalisability using the open-source \texttt{Qwen3.5-9B-DeepSeek-V4-Flash} model, achieving a PAR of 0.86 and a 54.5% latency reduction (to 10.2s). To evaluate practical utility and socio-technical integration, we administer detailed questionnaires to Ghanaian Extension Services Officers ($N=30$) and smallholder farmers ($N=36$). Taken together, this work demonstrates how policy-grounded structured retrieval-augmented generation with validated-memory routing makes safety-utility-latency trade-offs explicit, offering a scalable template for trustworthy AI-driven extension in smallholder farming systems.
arXiv:2606.14999v2 Announce Type: replace
Abstract: Scientific user facilities generate X-ray scattering data faster than traditional workflows can process them. We address this challenge across two settings, offline dataset exploration and live on-the-fly analysis. We train a domain-specific attention-based Convolutional Variational Autoencoder (C-VAE) on 1.5 million X-ray scattering images to learn low-dimensional representations capturing structural variation across diverse experimental conditions. The learned latent space reveals well-organized clusters and smooth trajectories reflecting experimental progression. It further supports controlled synthetic scattering image generation across diverse structural states. When deployed without retraining, the model organizes time-resolved film formation experiments at two synchrotron facilities into interpretable latent structures. Benchmarking against DINOv3 (ViT-7B), a general-purpose vision foundation model, demonstrates that domain-specific training yields more interpretable latent organization for scattering data. Both workflows are integrated within Latent Space Explorer, a component of the MLExchange platform, supporting interactive structural exploration across archived datasets and live experiments.
arXiv:2606.16776v3 Announce Type: replace
Abstract: Generalist robot policies require trustworthy evaluation and robot-usable training data, but both are difficult to scale with physical robots alone. Real-robot trials and demonstrations remain the most faithful source of deployment signals, yet they are slow, costly, and hard to reproduce. We present JoyAI-Sim, a simulation-enabled interconversion toolchain for human-robot aligned model evaluation and data generation, denoted as Robot $\rightleftharpoons$ Simulation $\rightleftharpoons$ Human. On the one hand, the Robot $\rightarrow$ Simulation $\rightarrow$ Human pathway supports human-robot aligned model evaluation by reconstructing real-robot tabletop organization tasks as calibrated digital twins for scalable evaluation, while using human embodied feedback to inspect and refine the naturalness of simulated motions. On the other hand, the Human $\rightarrow$ Simulation $\rightarrow$ Robot pathway supports human-robot aligned data generation: it lifts ego-centric human demonstrations into simulation, checks them under robot physical constraints, and converts them into robot-centered trajectories, annotations, and visual observations. Together, these pathways use the JoySim simulator as both a scalable evaluation layer and a physical consistency filter for robot data generation. We further package the core reconstruction, simulation, rendering, and realism-augmentation modules as cloud services on JD Cloud, turning the system into a reusable and scalable infrastructure for robot data generation and model evaluation.
arXiv:2606.20351v2 Announce Type: replace
Abstract: The standard convex-algebra interchange axiom, common to probability-monad formalisations since Stone, is provably too weak to support full Bayesian conditioning. We make this precise in Cubical Agda: finite distributions as a higher inductive type, conditional independence as a cubical path between kernels, recursive Bayesian conditioning as a total function on a full-support fragment. Lifting conditioning to the full HIT exposes a structural mismatch -- the two halves of the rearranged 4-leaf mix carry distinct Bayesian weights related by Bayes' formula, not the single shared inner weight the standard axiom provides. We exhibit the minimal generalisation that resolves this and prove the standard form is the degenerate case where the two inner weights coincide. Around this observation we verify the algebraic context constructively, with zero postulates above an abstract ordered-field interface: bind commutativity, the four semi-graphoid axioms, intersection (reduced to contraction via structural $\Sigma$-witnesses, without positivity), Pearl's do-calculus Rules~1, 2, and~3 in kernel form, finite-type Bayesian conditioning, and the soundness of Pearl's d-separation theorem on arbitrary finite directed acyclic graphs (DAGs) -- in interventional form for multi-element $X$, $Y$, $Z$, and in Bayesian form for the elementary patterns. The probability monad is also verified as a Markov category; the abstract interface discharges at $\mathbb{Q}$.
arXiv:2606.06223v2 Announce Type: replace
Abstract: Language-model agents act through repeated cycles of observation, reasoning, and action selection, making safety monitoring depend on both internal model state and environment context. We study reward-hacking monitors in ReAct-style agents acting in Gameable ALFWorld and WebShop. Agents are instrumented with activation-based reward-hack scores, token-level entropy, and decision-context features. We find that adapters fine-tuned on \textit{School-of-Reward-Hacks} dataset can transfer reward-hack tendencies into agentic action selection, especially when the environment exposes proxy-reward affordances. However, mitigating such behavior cannot rely on activation dynamics alone. High reward-hack activation identifies a latent policy state, but does not necessarily imply an immediate exploit action. Across next-step prediction tasks, entropy and context-calibrated internal features improve risk estimation over reward-hack activation alone. Activation-direction steering further reduces proxy-exploit behavior in selected mixed-adapter regimes. Overall, our results support context-calibrated internal monitoring for agents: reward-hack activation identifies a latent policy state, while entropy and decision context help determine when that state becomes risky action.
arXiv:2606.22804v3 Announce Type: replace
Abstract: Long, continuous video streams are an increasingly critical driver of multimedia intelligence. Existing efforts often handle long videos with a sample-encode-reason approach using large models. However, they overlook a crucial deployment fact: the stream is often produced by computationally constrained devices. This forces an untenable compromise: cloud offloading unlocks strong reasoning but incurs prohibitive bandwidth overhead, while on-device processing remains limited by edge hardware capacity. Therefore, we propose CoVStream, the first edge-cloud collaborative framework for understanding long video streams. The edge node distills raw video streams into compact visual features and semantic captions for transmission to the cloud, minimizing bandwidth costs, while the cloud server integrates this data into an entity graph and global visual context, activating the heavy reasoning model only when a user query arrives. Experiments on VideoMME-Long, LVBench, and RTV-Bench show that CoVStream reduces bandwidth usage by 87.6% while retaining 99.2% of the cloud baseline accuracy on LVBench.
arXiv:2606.24152v2 Announce Type: replace
Abstract: Large-scale video generation models are increasingly described as world models because they can learn rich spatiotemporal regularities from visual data. However, we argue that an ideal world model should benefit in a self-evolving generative character. Traditional visually plausible predictions alone are not enough to establish whether an imagined future is physically actionable for a particular embodied agent, failing to provide informative feedback from environments for self-evolving improvement. To realize self-evolving world models, this article proposes the concept of autonomous video generation, which is evaluated through counterfactual controllability, i.e., the ability to i) generate intervention-conditioned futures, ii) bind these future frames to embodiment constraints, iii) verify them under distribution shifts, and iv) distil surviving branches into compact variables for decision-making. We formalize a four-stage closed-loop optimization of Generation, Binding, Verification and Distillation, together with four corresponding evaluation metrics: novelty, consistency, out-of-distribution (OOD) and efficiency. We further discuss two examples, i.e., drones and manipulators, as early embodied testbeds where wind, sensing limits, actuation delay, contact dynamics and recovery constraints can be systematically perturbed and verified. The central claim is that the framework of autonomous video generation for self-evolving world models should not be judged by video fidelity alone, but by whether the generated frames improve valid action under counterfactual interventions and various embodiment constraints.
arXiv:2606.26580v2 Announce Type: replace
Abstract: Synthetic electrocardiogram (ECG) generation can support algorithm development and robustness evaluation, but simulated signals must preserve interpretable activation, recovery, and morphology properties. We present a graph-based ECG synthesis framework that combines activation-consistency certification with diagnostics-aware morphology curation. A unified heart graph supports an eikonal-template backend (ET) and a pseudo-diffusion reaction--eikonal backend (RE). We formulate graph Eikonal activation as a Bellman fixed-point problem and use the Bellman residual as a computable certificate for activation-time consistency. Each simulated ECG is evaluated by a two-stage diagnostics pipeline that separates metric computation from experiment-specific acceptance policies. On the cardiac graph, RE-derived activation times showed near-millisecond agreement with the Eikonal backbone and achieved $R^2=0.99876$ after causal predecessor filtering. Recovery experiments showed that endo-epicardial APD gradients determined the main T-wave morphology window, whereas the diffusion strength $\kappa$ provided secondary repolarization smoothing. In final balanced multi-lead curation, RE accepted 658/2000 samples versus 578/2000 for ET and increased per-model morphology coverage from 0.09248 to 0.09888. The framework provides a conservative basis for controllable and curated synthetic ECG generation.
arXiv:2606.27745v2 Announce Type: replace
Abstract: Panoramic images capture the full visual sphere in a frame, offering context unavailable to conventional cameras. Yet this completeness has an unavoidable geometric cost: the 2-sphere cannot be faithfully mapped to the plane, and every projection introduces distortions that challenge standard vision architectures. This survey traces panoramic scene understanding from projection-based adaptation and distortion-aware engineering to sphere-native modeling, reflecting increasing commitment to spherical geometry. Foundation models form a fourth family, geometry-aware tokenization, which adapts the input interface while reusing perspective-pretrained weights. We review these approaches across five task families: dense prediction, unified multi-task understanding, open-world perception, vision-language reasoning, and dynamic video analysis. Across tasks, the same shift toward spherical geometry recurs. In practice, however, the field has converged not on the strongest sphere-native operators, which are exactly rotation-equivariant but cannot reuse perspective-pretrained backbones and thus have not scaled, but on a compatibility-preserving middle ground combining moderate geometric awareness with large pretrained models. This commitment is uneven: deepest in dense prediction and shallowest in dynamic perception, where methods are spatially sphere-aware yet temporally planar. Foundation-model adaptation has advanced panoramic depth fastest, while layout, surface-normal, and video-level understanding remain largely unexplored. No panoramic foundation model has yet been pretrained on spherical data. We identify five evaluation gaps: spherical-area-weighted metrics, seam-consistency tests, polar-robustness stratification, cross-projection generalization, and standardized open-world protocols. We conclude with a six-point roadmap toward general-purpose panoramic intelligence.
arXiv:2606.28419v2 Announce Type: replace
Abstract: Limited data availability, class imbalance, and domain variability remain major barriers to reliable medical image classification. Conventional augmentation can improve training diversity but may distort diagnostically informative structures, whereas unconstrained generative augmentation may introduce label-inconsistent content. This paper proposes MedDiffuseMix, a saliency-guided diffusion mixing framework for controlled medical image augmentation. The method uses classifier-derived saliency maps to separate high-saliency diagnostic regions from low-saliency background areas and applies diffusion-guided mixing mainly to regions with lower diagnostic importance. Adaptive mixing, Gaussian boundary blending, and a saliency-preservation constraint reduce semantic distortion and reject or attenuate samples that shift model attention away from clinically relevant evidence. The framework is evaluated on four public benchmarks: the Radiological Society of North America pneumonia chest radiography dataset, Musculoskeletal Radiographs, PatchCamelyon, and the Breast Cancer Histopathological Image Classification dataset. Experiments with convolutional and transformer-based classifiers show that MedDiffuseMix improves accuracy, F1-score, and area under the receiver operating characteristic curve compared with standard augmentation, Mixup, GenMix, SaliencyMix, and diffusion-based augmentation baselines. Ablation studies confirm the importance of saliency guidance, adaptive region mixing, and smooth boundary blending. Visual attribution analysis further indicates that MedDiffuseMix better preserves diagnostically salient regions. These results suggest that saliency-guided diffusion mixing is an effective augmentation strategy for limited-data medical image classification.
arXiv:2606.29846v2 Announce Type: replace
Abstract: Shared autonomy systems combine user input with autonomous assistance to help users with motor impairments control robot arms to perform everyday manipulation tasks, by inferring user goals and providing appropriate guidance. However, the robot's internal beliefs about user goals cannot be observed by users. Traditional shared autonomy systems provide assistance along efficient shortest paths toward inferred goals, but when multiple objects lie in similar directions, such assistive motion remains ambiguous and fails to reveal the specific goal identified by the robot. This creates two critical problems. First, when the robot correctly infers the goal, users continue controlling because they cannot perceive understanding from ambiguous assistive motion, wasting effort when autonomous completion would suffice. Second, when the robot misunderstands intent, users cannot quickly detect errors until assistive motion diverges significantly, requiring substantial corrective input. We address this by introducing legible motion into shared autonomy, where robot actions must both advance toward the goal and clearly reveal which goal has been inferred, enabling users to understand the robot's beliefs and adjust control accordingly. The robot modulates communication strength through confidence-aware adaptive authority allocation by providing assertive legible assistive actions when confident while increasing user authority when uncertain, transforming shared autonomy into transparent bidirectional collaboration. User studies including simulation, and physical experiments with a six-degree-of-freedom robot arm demonstrate that legible shared autonomy significantly improves users' understanding of robot beliefs and reduces user control effort compared to standard shared autonomy.
arXiv:2606.29909v2 Announce Type: replace
Abstract: Encrypted traffic classification has achieved strong performance, but its decision process remains difficult to interpret. Existing methods usually combine flow statistics, packet sequences, and byte-level representations into opaque latent features, making it unclear which type of evidence actually drives the prediction. In this paper, we propose Traffic-CBM, a structurally interpretable multimodal framework for encrypted traffic classification. Instead of directly fusing heterogeneous traffic signals into a black-box representation, Traffic-CBM organizes them into a unified hierarchical concept space. These concepts are not manually annotated semantic attributes; rather, they are scalar evidence summaries constrained by predefined traffic evidence groups. More specifically, grouped flow statistics are mapped to statistical concepts, dedicated temporal encoders learn temporal concepts from disjoint feature subspaces, and byte-level evidence is further organized into packet-level and cross-packet concepts. This design turns heterogeneous traffic evidence into an explicit concept representation and makes different levels of traffic evidence easier to analyze. We evaluate Traffic-CBM on multiple encrypted traffic benchmarks. Results show that it achieves competitive and balanced classification performance while providing a clearer structural interpretation interface than conventional end-to-end fusion models. Further analyses suggest that the learned concept space is actively used in the prediction process and provides a clearer structural explanation of multimodal traffic evidence.
arXiv:2606.30248v2 Announce Type: replace
Abstract: Recent text-to-video (T2V) diffusion models rely heavily on auxiliary reward signals (e.g., via reward models or DPO) to align generated content with human aesthetics and improve realism. These signals, however, incur substantial computational overhead, require costly human annotations, and often yield limited improvement in fine-grained local details. In this paper, we argue that your data manifold is secretly a reward model. By explicitly modeling the manifold structure of high-quality Supervised Fine-Tuning (SFT) data and encouraging video latents to lie on this manifold, we derive dense, differentiable, and nearly cost-free reward signals that significantly improve video quality, particularly in mitigating low-level distortions. Our modeling builds upon Local Coordinate Coding (LCC), which captures the `skeleton' of the manifold. However, directly applying LCC suffers from mean regression, pulling latents toward the geometric mean and losing high-frequency details. We therefore extend it to Shell Local Coordinate Coding (Shell-LCC), which models the manifold `surface' as an isotropic shell to align with the true high-density region. Experiments demonstrate that our approach improves realism, enhances high-frequency details, reduces over-smoothing artifacts, and alleviates motion blur.
arXiv:2606.31043v3 Announce Type: replace
Abstract: Residual reinforcement learning adapts a pretrained robot policy by learning an additive correction to its actions. While effective when adaptation amounts to shifting the base policy's action distribution, additive corrections cannot change the distribution's shape, scale, or state-dependent geometry -- limitations we formalize as wrong variance, miscalibrated confidence, and non-uniform correction. We show that these matter under dynamics shift: when the base distribution is geometrically mismatched to the shifted system, residual correction can underperform even the unadapted policy. We propose Warp RL, a policy adaptation method that replaces additive residuals with an invertible, state-conditioned transformation of the base policy's action distribution. Instantiated with monotonic rational-quadratic spline flows (arXiv:1906.04032), Warp RL preserves identity initialization, strictly generalizes additive residual correction, and exposes a structured adaptation space suitable for both policy-gradient and gradient-free optimization. Across a variety of ManiSkill3 manipulation tasks with controlled dynamics shifts, Warp RL matches residual correction when translation is sufficient and substantially outperforms it when adaptation requires distributional reshaping. We further demonstrate that warping can replace additive correction in an off-policy sim-to-real pipeline, achieving comparable success rate with 30% faster task completion on a real-robot peg-insertion task.
arXiv:2606.31650v3 Announce Type: replace
Abstract: Long-horizon language agents must repeatedly interact with tools, accumulate evidence, and make decisions under bounded context windows. Context-management methods make such rollouts feasible by simplifying past interactions through deletion, folding, or memory editing. However, when useful history is collapsed into compressed states, the reconstructed context may no longer reveal which earlier observations support a successful final answer. This creates a mismatch between bounded-context acting and outcome-based reinforcement learning: the policy acts on reconstructed context, while the learner lacks source-level provenance for assigning credit to the evidence that mattered. We propose ECHO, a selective turn-memory framework for traceable context reconstruction in Agentic RL. ECHO compresses each completed environment turn into a compact source-indexed memory record, reconstructs bounded policy contexts by selecting useful records, and reuses the selected source indices to route positive outcome credit to the final trajectory segment, reused evidence turns, memory findings, and memory-selection actions. On BrowseComp-Plus, ECHO reaches 43.4% held-out accuracy, outperforming GRPO at 28.9% and the rolling-summary baseline SUPO at 36.1%, while using fewer turns and lower trajectory volume than SUPO. The trained policy also improves zero-shot generalization across multi-objective QA, code generation, and deep information-seeking benchmarks on both dense and MoE backbones.
arXiv:2607.00405v2 Announce Type: replace
Abstract: This paper presents a unified geometric, mathematical, and computational framework for the generation of the $complete$ admissible Pareto frontier. Several existing methods are structurally unable to capture the complete admissible Pareto frontier. These include widely used methods such as the weighted sum, the Normal Boundary Intersection (NBI) method, and the Normalized Normal Constraint (NNC) method. NNC and NBI, which share the same Pareto-generation grid construction, are structurally unable to capture 50% of the admissible Pareto region for tri-objective problems. More generally, for an $n$-objective problem, the admissible capture fraction decreases factorially as $1/(n-1)!$, and the corresponding missed fraction increases to $1-1/(n-1)!$. By contrast, the newly developed Generalized Normal Constraint (GNC) method introduced in the present work is structurally capable of capturing the complete admissible Pareto frontier. The proposed GNC method is formulated for general $n$-objective optimization problems and is developed through a unified geometric, mathematical, and computational framework supported by computational examples. Multiobjective optimization plays an important role in a broad range of applications, including economics, product design, and engineering management. Accordingly, the ability of a Pareto-generation method to generate a representative subset spanning the $complete$ admissible Pareto frontier is of fundamental importance for multiobjective optimization.
arXiv:2607.00867v3 Announce Type: replace
Abstract: Long-video reasoning is fundamentally constrained by how models acquire and utilize visual evidence. Existing tool-augmented video frameworks often interleave temporal grounding and answer reasoning within a single trajectory, causing early semantic hypotheses to bias evidence localization. We term this failure mode premature semantic commitment, where biased grounding retrieves incomplete evidence and incomplete evidence further reinforces incorrect reasoning. To address this issue, we propose EFlow, an evidence-first video reasoning framework built upon Qwen3-VL. EFlow explicitly separates temporal grounding and logical reasoning through CoT for Temporal Grounding and CoT for Reasoning, enabling the model to retrieve relevant evidence before answer inference. In addition, EFlow introduces a confidence-aware reflection mechanism that re-evaluates the full video when retrieved evidence is potentially insufficient. We further construct dedicated trajectory datasets and train EFlow through supervised fine-tuning, reinforcement learning, and reinforcement fine-tuning. Extensive experiments across five video understanding benchmarks demonstrate that EFlow consistently improves long-video reasoning performance.
arXiv:2607.01060v4 Announce Type: replace
Abstract: Video world models are emerging as a scalable alternative for evaluating generalist robot policies, bypassing the physical constraints and engineering burdens of real-world deployment. However, evaluating policies with video world models remains challenging, as world-model errors can make generated rollouts unreliable and slow inference limits large-scale throughput. We introduce RoboWorld, an automated evaluation pipeline that pairs a fast autoregressive video world model with a task-progress-aware vision-language model scoring. To enable reliable long-horizon autoregressive world-model rollouts, we propose Step Forcing, which combines anchored and one-step self-forwarded contexts to reduce train-test mismatch while preserving action-observation dynamics. Together, these components enable RoboWorld to align strongly with real-world robot evaluation across tasks and environments, achieving Pearson's r = 0.989 and Spearman's $\rho$ = 0.970.
arXiv:2607.01435v2 Announce Type: replace
Abstract: Significant advancement of immersive technologies such as Virtual and Augmented Reality (VR/AR) and their integration into diverse aspects of modern life need authentication interfaces that are secure, intuitive, and compatible with embodied interaction. Traditional methods such as passwords, PINs, and device-based logins, break immersion and rely on external hardware. Recent 3D-specific behavioral approaches, such as hand-gesture, eye-tracking, and electroencephalography (EEG)-based methods, offer promising alternatives but often require specialized sensors or constrain natural movement, limiting usability in dynamic environments. We present Sign in the Air to Unlock, an in-air signature interface that enables users to authenticate by signing naturally in 3D space which is a familiar, personal, and reproducible gesture. To realize this interface, we design a point-voxel Cross-Attention Network (PV-Net) that jointly models local motion dynamics and global spatial structure from 3D trajectories. The model is evaluated on two datasets: the public DeepAirSig dataset (1,800 signatures from 40 users) and ImmAirsig, a new dataset collected using Meta Quest 2 in immersive VR (880 samples from 22 users). PV-Net achieves an Equal Error Rate of 2.5% on DeepAirSig and 76% classification accuracy on ImmAirSig. These findings highlight the potential of 3D behavioral interfaces for seamless, user-centric authentication that merges security with natural interaction in immersive environments.
arXiv:2607.01455v3 Announce Type: replace
Abstract: Language models learn continuous programs over discrete symbols, with the embedding table and LM-head acting as the read/write interface between them. We show that this interface has gradient geometry distinct from dense hidden weights which can be exploited to improve the Pareto frontier across supervised finetuning, RL, and pretraining, while only utilizing kilobytes of optimizer state. We introduce Ember, a lightweight optimizer for embedding and LM-head matrices that utilizes O(V + D) VRAM, instead of Adam's O(2VD), and forgoes the need to shard both token table optimizer states. We provide empirical evidence that Ember scales effectively across batch size and parameter count. We show that the optimization trajectory of tokens can be well described by a simple 1D ray, counter to the popular belief that neural net parameters navigate a heavily nonconvex landscape. We provide a principled view on the surprisingly narrow space of optimizers that suffice for Transformer training. Finally, we open-source our distributed Ember implementation that merges cleanly with existing ZeRO/FSDP setups to support further research at https://github.com/katop1234/ember