arXiv subjects

Lin Wang

Publications and source records attributed to Lin Wang.

At least 19 recordsLinked to original sources

arXiv2026

From Intent to Execution Grant: An Execution-Boundary Conformance Profile for High-Risk AI Actions

AI agents increasingly propose actions with external consequences, including financial transfers, infrastructure changes, software deployments, disclosures, and physical actuation. Authorization engines, policy languages, runtime monitors, provenance mechanisms, and agent guardrails provide important foundations, but do not necessarily define a common semantic contract for the final transition from a particular candidate action to execution authority. We specify EBL-Core, an execution-boundary conformance profile for deciding whether one canonical, fully materialized AI-generated candidate may receive action-scoped execution authority under explicit conditions. It binds a structured intent object, Root and Operational Policies, evidence obligations, typed evidence, context, time, and a verifiable Decision Derivation through an Execution Release Contract (ERC). An ERC is not an authority-bearing token; a verified ALLOW ERC may support a separate Execution Grant governed by Redemption-time validation. EBL-Core specifies action binding, policy non-weakening, evidence handling, deterministic adjudication, derivation verification, and grant lifecycle behavior. An accompanying reference artifact provides schemas, adjudication, separate verification and Semantic Replay, and a linearizable in-memory grant store. In the retained run, 34 static vectors and 15 lifecycle checks matched expected outcomes. Across 100 trials, 32 concurrent Redemption attempts yielded exactly one successful Redemption and protected test effect per trial; 100 Revoke-Redeem races ended in valid terminal outcomes. These bounded results demonstrate executability of the specified subset, not human-intent correctness, evidence truth, complete mediation, production readiness, mechanized correctness, or deployment-level security.

cs.CR↗

arXiv2026

Morphological Decoupling-Based Skeletal Classification for Clinical Assessment of Malocclusion

Malocclusion skeletal grading is a fundamental task in orthodontics, critical for diagnosis and treatment planning. Traditionally, cone-beam computed tomography (CBCT) is used for visual measurement, and the reconstructed lateral cephalograms are handed over to expert dentists for diagnosis. However, manual review is time-consuming, labor-intensive, and subject to inter-operator variability. Therefore, an automatic CBCT-based system is needed for reliable malocclusion skeletal grading. In this case, we develop TeethGNN, a novel graph-based framework designed to combine CBCT image features with morphological information for accurate and efficient malocclusion grading. TeethGNN utilizes a decoupled learnable decoder to directly predict key morphological indicators from CBCT images, eliminating the need for manual measurements. These morphological features are then fused with image features using a graph neural network (GNN), which effectively models the relationships between the modalities. To further enhance robustness and calibration, we introduce a collaborative calibration strategy. This strategy combines multi-scale graph adversarial perturbation for explicit calibration and nonlinear topological graph calibration for implicit confidence adjustment. Extensive experiments and ablation studies on our collected clinical dataset demonstrate that our malocclusion measurement system achieves 77.08\% in accuracy and 89.61\% in AUC, outperforming the compared state-of-the-art methods. These results validate the effectiveness of graph-based multimodal fusion and collaborative calibration in improving malocclusion grading performance. Our system shows strong potential for advancing computer-aided orthodontic diagnosis, providing an accurate and reliable solution for vision-based clinical measurement and diagnosis.

eess.IV↗

arXiv2026

CineForge: Self-Improving Agents for Long-Horizon Video Generation

Long-horizon story-driven video generation requires a production agent to coordinate narrative decomposition, state tracking, shot design, prompt construction, rendering, and revision across interdependent scenes. Existing adaptive video systems primarily refine requests or reusable skills, leaving recurring production failures disconnected from persistent, stage-targeted improvements across stories. We introduce CineForge, a self-evolving video-production agent framework that couples CineForge-Produce for video generation with CineForge-Evolve for cross-story policy evolution. CineForge-Produce organizes each source story into typed narrative, character, spatial, and cinematic states, uses them to coordinate asset and clip generation, and records the process as a canonical production trajectory. CineForge-Evolve applies Case-to-Pattern-to-Policy Evolution (CPPE) to review trajectory evidence, consolidate recurrent findings into bounded stage-local patches, and deploy validated updates through structural replay and confidence-controlled paired evaluation. To measure complete story realization, we introduce CineScope, which combines a 100-script CineScope-Data suite with a human-aligned, multiscale CineScope-Metric spanning causal state, directorial orchestration, pacing and resource allocation, and character arc. Across CineScope-Data and two public benchmarks, the evolved CineForge policy improves CineScope-Metric from 4.024 to 4.380, outperforms three long-video baselines with consistent gains under ScriptAgent, and reduces review LLM calls by 37.0% on new stories. These results establish production trajectories as actionable experience for video agents that improve cumulatively across long-form storytelling tasks.

cs.CV↗

arXiv2026

Resilience Matters for Embodied Agents System: New Metrics, Systematic Evaluation, and Optimization

Embodied Agents System (EAS) are increasingly deployed in open-world physical domains, where reliability directly dictates deployment quality and human-agent trust. However, existing evaluations rely on outcome-centric metrics as success rate or safety scores that collapse diverse execution trajectories into coarse scores, obscuring the dynamic processes underlying agent behavior. Therefore, they ignore a critical property of EAS -- which we define as the Resilience -- that reflects how EASs recover, stabilize, and extend under perturbations and across iterative updates. The lack of resilience is particularly critical in open-world environments due to continuous unexpected disruptions, thus directly affecting the quality of EAS deployment. To address this problem, we gain insight from the resilience-engineering concepts to EAS groundings and propose a novel resilience evaluation framework that can be flexibly applied to any EAS. Specifically, we define the first comprehensive resilience metrics suite for EASs system that exposes Rebound, Stability, and Graceful Extensibility across embodied tasks execution, providing a practical grounding for EAS resilience analysis. We further implement the resilience evaluation layer that transforms execution process into assessments for diagnosis and optimization. Across 400 household tasks with 10 EAS, we reveal the process-level distinction hidden by outcome metrics, including recovery cost differences among successful episodes ($\Delta C_{rec}=25.2$), increased instability and task-family degradation. Metrics-guided optimizations reduce recovery cost and increase stability, graceful extensibility completion, showing the diagnostic effect of resilience evaluation. Our results reveal a trade-off among resilience characteristics, suggesting that a resilient EAS construction should be configured according to deployment-specific requirements.

cs.RO↗

arXiv2026

Who Can Make the Action Happen? An Authority-Decomposition Framework for High-Risk Automated Systems

High-risk automated systems distribute control across services, credentials, protected components, and lifecycle mechanisms. Labels such as authorized, approved, privileged, or protected therefore do not answer a basic causal question: which actors can actually make a consequential action occur? This paper provides an action-relative method for deriving which trust-domain coalitions are sufficient to cause protected execution, defined as the occurrence of a designated protected state transition. The framework models components, powers, resources, boundaries, and alternative realization structures; includes update, recovery, override, disablement, and alternative invocation; and separates causal control over execution from control over the authoritative account of an operation. It derives inclusion-minimal sufficient coalitions and tests whether claimed execution boundaries remain independent of designated upstream domains. Cross-domain analytical cases illustrate the method. In a split-control, release-intended, open-state, source-bounded Havenlon protocol model, the ordinary witness requires five trust domains, while certificate replacement yields a three-domain inclusion-minimal known requirement set among source-enumerated protocol witnesses; the Linux domain remains insufficient for the complete transition. Deployed global non-bypassability and boundary-bound veto coverage remain unresolved. The framework is a conceptual and analytical tool. It does not certify implementations, establish deployment security, guarantee complete discovery of hidden powers, or define evidence-verification semantics.

cs.CR↗

arXiv2026

Energy-Structured Latent World Models with Neural Time Fields for Physically Constistent Open-World Motion Planning

Physically consistent motion planning remains a fundamental challenge in embodied AI, as generated trajectories must strictly conform to real-world execution dynamics. While latent world models offer a promising approach by predicting these dynamics, existing methods learn unconstrained future representations where absorbed physics remains implicit. Therefore, they fail to form reusable physical knowledge, which compromises reliability in unpredictable open-world navigation. To address this, we propose a novel Energy-Structured Latent World Model (ELWM). Our key idea is to structure the ELWM latent state to explicitly carry energy and momentum, ensuring strictly causal transitions via dissipation and control ports. Trained on multimodal RGB-D and inertial interaction histories, our model guarantees physically consistent predictions. We further implement this for motion planning by constructing Physics-Conditioned Neural Time Fields (PC-NTF), a key technical cornerstone that integrates ELWM into an arrival time field via the Eikonal equation to yield a physically-informed navigation policy. Across held-out scenes, our evaluation reveals significant improvements. Compared to generic latent models, PC-NTF reduces 0.8-s motion-prediction NRMSE from 0.36 to 0.29. Against Active Neural Time Fields, it improves navigation success from 81.3% to 89.7% and SPL from 0.64 to 0.73, while cutting the physical collision rate from 12.1% to 5.8% and the Eikonal residual from 0.083 to 0.031. Beyond these targeted gains, our results demonstrate that embedding explicit physical structures into latent spaces intrinsically bridges the gap between predictive world models and safe, dynamically feasible motion planning.

cs.RO↗

arXiv2026

On the sharpness of the $C^1$-norm threshold for perturbations in the normally hyperbolic invariant manifold theorem---a toy model perspective

The classical normally hyperbolic invariant manifold theorem asserts that a $C^1$ normally hyperbolic invariant manifold persists under $C^1$ small perturbations. For a family of standard-like dissipative twist maps, we show that the threshold $(1-\sqrt{\lambda})^2$ for the $C^1$-norm of the perturbation is sharp: there exists a $C^\infty$ perturbation $\phi$ with $\|\phi\|_{C^1} = (1-\sqrt{\lambda})^2$ such that the map preserves a unique invariant graph, but this graph possesses non-differentiable points. On the other hand, whenever $\|\phi\|_{C^1} < (1-\sqrt{\lambda})^2$, the $C^1$ normally hyperbolic invariant manifold persists, where $\lambda$ denotes the Jacobian determinant of the map. This provides a critical threshold phenomenon for the persistence of invariant graphs in dissipative twist maps.

math.DS↗

arXiv2026

Persistence of invariant graphs for twist maps under analytic perturbations

We consider the persistence for invariant graphs of twist maps that exhibit the strongest possible dynamics, namely those real-analytically conjugate to rigid rotations, under Gevrey-$\gamma$ ($\gamma\in [0,1]$) perturbations. By enhancing the regularity of the perturbation itself, we show that invariant graphs with the strongest dynamics can persist even when the size of the perturbation and the constraints on the frequency go beyond the requirements of classical KAM theory and the theory of normally hyperbolic invariant manifolds. The proofs of these results are based on a parameterized direct KAM method.

math.DS↗

arXiv2026

TSDM: A Scheduling Policy for Joint Throughput-AoI Optimization in Multichannel Wireless Networks

Optimizing for both low Age of Information (AoI) and high throughput is critical for remote sensing applications that rely on multichannel wireless networks. However, jointly optimizing these two metrics is an analytically challenging problem, particularly in systems with heterogeneous and unreliable channels. To address this challenge, we propose TSDM, a Two-Stage Deficit Matching scheduling framework. TSDM is based on a second-order approach that characterizes the performance of each data flow by its mean and temporal variance. In the first stage, TSDM translates the high-level utility maximization objective into a concrete set of target mean and temporal variance statistics for transmissions over each node-channel pair. In the second stage, a low-complexity Weighted Matching Deficit (WMD) rule performs real-time channel assignment. We theoretically prove that TSDM achieves the desired mean and temporal variance for each flow. Furthermore, we conduct extensive simulations on two open joint throughput-AoI optimization problems. In both cases, TSDM significantly outperforms existing scheduling policies.

eess.SY↗

arXiv2026

DMSNet: Cross-Band Learning for Multi-Target Sensing in Multi-Band ISAC

Multi-band integrated sensing and communication (ISAC) offers complementary high- and low-frequency echo information for multi-target sensing. However, existing dual-band ISAC sensing methods have a limited ability to exploit deep complementary information across heterogeneous bands and often incur high computational costs. To address these limitations, we propose a Dual-Band Multi-Target Sensing Neural Network (DMSNet) for joint target number and parameter estimation. Under representative simulation conditions, DMSNet outperforms the best baseline in target number estimation, increasing count accuracy from 89.01 % to 91.74 % and Macro-F1 from 90.80 % to 93.07 %. For parameter estimation, compared with the best baselines, DMSNet reduces the median absolute errors of range, velocity, and angle by 82.2%, 56.9%, and 73.2%, respectively. Moreover, DMSNet reduces runtime by 68.7 % relative to the fastest existing dual-band ISAC sensing method.

eess.SP↗

arXiv2026

SelectInfer: Selective Neuron Loading and Computation for On-Device LLMs

Large Language Models (LLMs) have demonstrated remarkable capabilities across a range of Natural Language Processing (NLP) tasks, but their high computational and memory demands pose significant challenges for deployment on resource-constrained edge devices. Existing approaches to model compression and optimization often rely on coarse-grained pruning or quantization, which can compromise accuracy or require re-training and fine-tuning. In this work, we introduce SelectInfer, a neuron-level optimization framework that enables efficient LLM inference on edge devices through selective neuron loading and computation. By profiling and identifying both task-specific and general-purpose neurons using an offline LLM profiler, SelectInfer implements two key optimizations: selective loading, which reduces memory footprint by selectively loading a subset of neurons that were identified to be most important during the offline stage, and selective computation, which dynamically computes only the most relevant neurons at runtime. Evaluation across multiple datasets shows that SelectInfer achieves significant reductions in memory footprint and computation while preserving task performance, making it a practical step towards enabling LLM deployment on edge devices

cs.LG↗

arXiv2026

VistaVLA: Geometry- and Semantic-Aware 3D Gaussian-Grounded VLA for Robotic Manipulation

Vision-Language-Action (VLA) models have emerged as a powerful end-to-end paradigm for robotic manipulation by mapping language instructions and 2D visual inputs directly to actions. However, these models lack an explicit, scene-level 3D representation, limiting their ability to reason over spatial layouts and geometric constraints. While recent efforts incorporate explicit 3D cues, such as depth maps or point clouds, to improve geometric awareness, they primarily capture low-level structures and lack high-level semantic grounding in 3D space. In human cognition, interaction with the physical world relies on a 3D semantic cognitive map - an internal mental model that integrates spatial layouts with semantic context to enable persistent, viewpoint-invariant reasoning. In light of this, we present VistaVLA, a novel two-stage framework that constructs a geometry- and semantics-aware 3D cognitive representation from 3D Gaussian primitives and grounds it as compact context tokens for VLA policy learning. Specifically, VistaVLA lifts multi-view vision-language features into 3D Gaussian primitives, forming geometry-anchored semantic tokens that align view-consistent spatial grounding with 2D visual feature spaces. To make this 3D representation computationally tractable for effective VLA control, we introduce Merge-then-Query (MtQ), a token summarization mechanism. MtQ compresses dense Gaussian primitives into a highly compact set of spatially informative tokens, achieving a 99% token reduction while preserving action-relevant 3D layouts and semantic context. Extensive evaluations in both simulated and real-world environments demonstrate the effectiveness of VistaVLA. Notably, in real-world scenarios, VistaVLA improves success rates by 22.8% across seven real-world tasks and by 30.0% over the VLA-Adapter baseline on challenging out-of-distribution tasks.

cs.RO↗

arXiv2026

UMSS: Towards Unsupervised Multi-modal Semantic Segmentation

Multimodal semantic segmentation (MSS) is essential for robust perception in complex environments, yet its potential remains largely untapped because of the prohibitive cost of human annotations. While unsupervised semantic segmentation (USS) has achieved strong results on a single RGB modality, its naive extension to multimodal data is often hindered by fusion degradation. This occurs because, without explicit supervision, existing frameworks struggle to reconcile the heterogeneous structural patterns captured by different sensors and therefore fail to effectively exploit their complementary information. In this paper, we make the first attempt to address the novel problem of Unsupervised Multimodal Semantic Segmentation (UMSS), aiming to effectively exploit complementary sensor information in a fully label free setting. To this end, we propose UniM2 (Unified Multimodal), a novel framework built on DINOv3 that transforms conventional fusion methods into consistent performance gains. Our key idea is to learn a unified latent space driven by Cross Modal Correspondence Synergy (CMCS) to extract intrinsic shared semantic cues, bypassing the need for label guided adaptive fusion. To mitigate inherent intermodal conflicts, we introduce a Cross Modal Harmonizer (CMH) that designates RGB as a stable reference, effectively suppressing inconsistent relational supervision while guiding the model to exploit complementary structural features. Extensive experimental results on NYU Depth v2 and MFNet show that UniM2 improves mIoU by 6.4% and 9.8%, respectively, demonstrating clear advantages over existing frameworks for UMSS.

cs.CV↗

arXiv2026

Physics-Guided Biomechanical Gait Adaptation for Humanoid Locomotion on Extreme Sloped Terrains

Model-free reinforcement learning has enabled impressive humanoid locomotion; however, control on steep slopes remains largely unexplored. Unlike flat or discrete terrains, sloped terrains impose a persistent gravitational bias that demands simultaneous stability and posture control. Consequently, under generic reward formulations, policies can converge to slow, conservative low-center-of-mass (CoM) crouched gaits. In this work, we propose a novel two-stage physics-guided framework, dubbed HumoSlope, dedicated to robust humanoid locomotion on diverse sloped terrains. Specifically, Stage I establishes a terrain-consistent balance prior by introducing a slope-adaptive Zero Moment Point (ZMP) regularizer evaluated directly on the local inclined support plane rather than a world-horizontal reference. To prevent the resulting policy from defaulting to a crouched posture, Stage II introduces the Biomechanical Slope Gait Adapter (BSGA). Utilizing extracted macroscopic terrain descriptors as privileged, training-only signals, BSGA dynamically gates soft reward priors to modulate CoM height and lower-limb coordination based on the estimated slope geometry -- encouraging hip-dominant uphill propulsion and knee-oriented downhill braking. Crucially, the deployed actor remains entirely proprioceptive, requiring no online exteroceptive sensing. Extensive Sim-to-Real experiments demonstrate that our framework effectively mitigates posture degeneration and enables blind, continuous traversal of outdoor grass slopes up to 62.7% ($32.1^\circ$), validating a physics-guided approach to challenging slope terrain adaptation.

cs.RO↗

arXiv2026

NKI-Agent: Domain-Specific Fine-Tuning and Agentic Tool Use for Neuron Kernel Generation

Recent agentic approaches to LLM-based kernel generation have achieved impressive results on CUDA. For emerging AI accelerators such as AWS Trainium and Inferentia, automated kernel generation and optimization remain largely unaddressed. Writing kernels for these chips via the Neuron Kernel Interface (NKI) is particularly challenging: developers must navigate a multi-engine architecture, tile-based programming, and explicit data movement across multi-level memory hierarchy. Moreover, no publicly-available training data, benchmarks, or tool-augmented agents exist for this domain. We introduce NKI-Agent, the first system combining domain-specific supervised fine-tuning (SFT) with a compile-verify-fix agent loop for NKI kernel generation. We adapt the existing CUDA-Agent framework to Neuron hardware, curate 6,000 NKI kernel generation tasks for training, and construct NKIBench, a 250-task benchmark across three difficulty levels. Evaluated on real Trn1 hardware, NKI-Agent with Claude Opus 4.8 and a rank-aware system prompt achieves a 77.3% pass rate on the 150-task NKIBench. We show that tool use is critical: Opus 4.8 scores 6% in single-shot mode without agent tools. On a 60-task subset, we show that an SFT-trained Qwen3-Coder-30B-A3B achieves 25.0% pass rate at 1/100th the cost, outperforming Claude Sonnet 4 (15.0%). We also report that Group Relative Policy Optimization (GRPO) with binary compilation reward fails to improve over SFT, providing guidance on reward design for RL-based kernel generation.

cs.LG↗

arXiv2026

LiMoDE: Rethinking Lifelong Robot Manipulation from a Mixture-of-Dynamic-Experts Perspective

Building a generalist robot that can leverage prior knowledge for continuous task adaptation remains a significant challenge. Previous works alleviate the catastrophic forgetting problem by parameter-efficient fine-tuning for single-task adaptation. However, they fail to extract reusable skills and model the interaction with other skills effectively. Recent works try to address these issues by learning prompts. Differently, this paper presents an architectural perspective on the Lifelong Mixture of Dynamic Experts (\textit{LiMoDE}), a novel two-stage learning scheme for lifelong robot manipulation. Specifically, a dynamic MoE structure is first proposed in the multi-task pre-training stage to learn prior knowledge, where a varied number of heterogeneous experts are activated based on the motion information to address different short-term manipulations. Subsequently, in the task adaptation stage, we design a lifelong MoE adaptation mechanism % (LiMoEAM) that learns lifelong experts and dynamically combines them with frozen ones for new tasks, facilitating the knowledge transfer during adaptation. The proposed \textit{LiMoDE} is evaluated on both the simulated lifelong learning benchmark and real-world tasks. Extensive experiments demonstrate its effectiveness in achieving superior performance and strong lifelong adaptation by introducing a moderate number of additional trainable parameters and inference overhead.

cs.RO↗

arXiv2026

A Systematic Survey on Event Camera Representation Learning

Event cameras offer distinctive advantages, including microsecond-level latency and high dynamic range, rendering them promising for challenging perception tasks. Inspired by biological vision, they output asynchronous and sparse event streams rather than dense image frames, creating a fundamental mismatch with mainstream neural networks. This survey reviews recent advances in event camera representation learning from the perspective of converting raw event streams into learnable representations. We first organize existing methods according to whether they rely on a single principal event representation or jointly exploit multiple complementary representations. Single-representation methods are further categorized into dense-based representations, which regularize events into structured grid-like forms, and sparse-based representations, which preserve event-native discrete spatio-temporal structures. Multi-representation methods are organized into dense-dense and dense-sparse hybrid formulations that exploit complementary representation properties. This representation-centric taxonomy clarifies how different paradigms balance structural regularity, temporal fidelity, sparsity preservation, and architectural compatibility. For each paradigm, we examine the underlying design choices, modeling principles, and task-level implications. We further summarize standard benchmarks and evaluation settings across representative high-level perception and low-level vision tasks. Finally, we discuss open problems and outline future directions from fixed representation design toward adaptive representation optimization, improved fidelity-efficiency trade-offs, and more scalable event-based perception systems.

eess.IV↗

arXiv2026

A Whisper from Within: Response of a Pulsar Timing Array to an Internal Gravitational-wave Source

Millisecond pulsars (MSPs) are abundant in globular clusters (GCs) and probably also in galactic nuclei. They offer the potential to form a miniature pulsar timing array (mini-PTA) to detect nanohertz gravitational-wave (GW) sources located inside the array. Since the size of such an array is comparable to the wavelength of GW, the conventional plane-wave approximation becomes invalid, and near-field effects, including wavefront curvature, non-radiative self-field of the GW source, and direct perturbation of pulsar by GW, become significant. In this work, we incorporate these effects in a comprehensive model to calculate the timing residual induced by a GW source inside a mini-PTA. We also consider realistic GW source configurations in GCs (M15 and $\omega$ Centauri) and in galactic nuclei (Sgr A* and M31), and find that for MSPs located sufficiently close to the GW source (within a few wavelengths), the residual can reach $1~\mu\mathrm{s}$ in GCs and up to milliseconds in galactic centers, within the potential detection reach of current radio telescopes. Crucially, when the pulsar lies within a few GW wavelengths of the source, the non-radiative field dominates and causes the residual to rise much more steeply (between $1/r_e^2$ and $1/r_e^4$, where $r_e$ is the distance to the source) than the conventional far-field scaling ($1/r_e$). These results demonstrate that mini-PTAs in GCs or galactic nuclei can serve as powerful probes of otherwise invisible GW sources, including intermediate-mass and supermassive black hole binaries.

astro-ph.HE↗