Searcharxiv⌕ Search

arXiv subjects

Rui Jiang

Publications and source records attributed to Rui Jiang.

At least 19 recordsLinked to original sources
arXiv2026

EmbodiExp: Compositional Experience Transfer for Embodied Agents

Experience can help embodied agents turn high-level plans into grounded actions, but transferring it across tasks remains challenging: previously effective strategies may be inapplicable to the current state or fail to produce their intended outcomes. We present EmbodiExp, a framework for constructing transferable experiences and composing them online during embodied task execution. Our key observation is that tasks often share subgoals whose completion can be assessed independently of overall task success. We therefore introduce an outcome-grounded subgoal experience that specifies a subgoal, applicability conditions, an adaptable policy, and a verifiable outcome. During execution, applicable experiences guide action generation, while verified outcomes can establish the conditions required by subsequent experiences, enabling online composition. Subgoal-level outcome attribution further enables failure utilization: failed trajectories can still contribute verified local successes and condition-specific guidance on actions to avoid. Experiments on multiple embodied benchmarks under a fold-based protocol with frozen experience banks demonstrate the effectiveness of EmbodiExp.

cs.RO↗
arXiv2026

UniRect-CoT: Enhancing Generation in Unified Multimodal Models via Reflective Rectification with Inherent Understanding

Unified Multimodal Models (UMMs) aim to integrate visual understanding and generation within a single structure. However, these models exhibit a notable capability mismatch, where their understanding capability often outperforms their generation capability. This mismatch suggests that the model's rich internal knowledge, while effective for understanding tasks, is not fully utilized during generation. To address this, we draw inspiration from the human ``Thinking-While-Drawing'' paradigm, where humans continuously reflect on intermediate results and revise them according to their understanding of the intended target. In this paper, we propose UniRect-CoT, a training-free unified rectification chain-of-thought framework. We regard the multi-step denoising process in UMMs as an intrinsic visual reasoning process, whose intermediate states provide opportunities for continuous reflection and rectification. By leveraging the UMM's inherent understanding to reflect on intermediate results in light of the target instruction and rectify them accordingly, our framework forms a reflective chain of thought within a single denoising trajectory, enabling the model's internal knowledge to guide generation throughout the process. Specifically, the UMM generates a state-conditioned reflection, whose autoregressive negative log-likelihood under the same model defines a model-native differentiable semantic loss that translates textual reflection into gradients for latent rectification. Extensive experiments demonstrate that UniRect-CoT generalizes across existing flow-based UMMs, consistently enhancing overall text-to-image generation performance.

cs.CV↗
arXiv2026

PoVD: Efficient Consensus Protocol based on Verifiable Delay Function

Consensus protocols ensure the robustness and scalability of blockchains and decentralized applications built on them. However, existing consensus mechanisms often impose high computational cost or require heavy communication overhead. To address these challenges, we propose proof of verifiable delay (PoVD), a lightweight consensus protocol based on the verifiable delay function (VDF). We present the detailed protocol of PoVD, including the block mining and verification rules, and illustrate how PoVD can flexibly adjust the block time distribution according to the network condition. Through mathematical proof, we point out that, under a relatively weak network assumption, PoVD can achieve a lower fork rate than PoW while maintaining the same throughput. Our experiments verify that PoVD exhibits low computation and communication complexity and can also improve the blockchain consistency, making it particularly suitable for resource-constrained nodes and bandwidth-limited networks.

cs.DC↗
arXiv2026

A Strictly Proper Scoring-Rule Theory for Calibrating Stochastic Car-Following Models

Problem definition: Fixed parameters and inputs in a stochastic simulator induce a distribution over complete trajectories, not one trajectory. Calibration must assess this distribution, including variability and temporal dependence, against observations. Yet stochastic car-following models are commonly calibrated with trajectory-error objectives inherited from deterministic modelling. Methodology/results: We establish a scoring-rule theory of stochastic calibration. Strict propriety requires the data-generating distribution to uniquely minimise expected score. MRMean-I, the average run-wise error, drives separable stochastic spread to zero; MRMean-II, the error of the ensemble-mean trajectory, cannot identify a parameter that changes only spread; and MRMin, the error of the closest simulated run, has a population target that changes with ensemble size. These results are confirmed for stochastic Intelligent Driver Model extensions with additive acceleration noise and random desired headway. We recommend exact maximum likelihood when the correct transition density is available; otherwise, an unbiased simulation-based estimator of a strictly proper score. The energy score meets this requirement and gives the best held-out distributional prediction among the evaluated simulation-based objectives, although both models retain too-narrow bands and miss persistent disturbances. Implications:Strict propriety separates a valid calibration target from parameter identifiability and model adequacy. The theory applies to vector-valued outputs from stochastic transportation simulators; the car-following experiments illustrate its scope.

stat.ME↗
arXiv2026

Embedded Graph Flows for Categorical Graph Generation

Generating categorical graphs requires choosing node and edge types that form a coherent structure without depending on node order. Many graph generators encode categories as fixed one-hot vectors, which can impose an artificial geometry in which categories are equidistant. We propose Embedded Graph Flows (EGF), a generative model that learns continuous embeddings for node and unordered-edge categories and transports Gaussian noise towards these learnt endpoints using a permutation-equivariant graph transformer. A terminal readout maps the embeddings back to discrete graph categories. Across molecular benchmarks, EGF achieved competitive performance. On QM9, EGF gives the best result on all four reported metrics among the three methods, including a Fréchet ChemNet Distance (FCD) of 0.150, compared with 0.717 for the categorical-diffusion baseline DiGress and 0.812 for the bridge-based baseline GruM. When applied to larger molecules in ZINC250k, EGF retains the lowest maximum mean discrepancy (MMD) using the neighbourhood subgraph pairwise distance kernel (NSPDK), indicating close agreement with the local substructures of the reference molecules. Our code is available at https://github.com/Trusted-System-Lab/EGF.

cs.LG↗
arXiv2026

Evaluating Investment Logic in Large Language Models: A Real-World Benchmark Towards Personalzied Financial Agents

Investment competence is inherently personalized: the same market evidence can justify different actions for investors with different goals, horizons, portfolios, and risk boundaries. Yet financial LLMs are evaluated either by static question answering or by terminal profit and loss. The former omits agency; the latter cannot reveal whether a profitable action was grounded, profile-consistent, or merely lucky. We ask whether the community is using the wrong ruler for consequential agents. We introduce \textsc{InvestLogicBench}, a process-native benchmark containing 201,247 documented decisions from 151 real-world investors. Each episode instantiates a \textbf{P$\rightarrow$E$\rightarrow$R$\rightarrow$D$\rightarrow$O} trace: investor \textit{Profile}, observable market \textit{Events}, investment \textit{Reasoning}, executable \textit{Decision}, and delayed \textit{Outcome}. The release includes profile construction, point-in-time event binding, structured logic, horizons, outcomes, and post-mortems, and supports comprehension, profile-conditioned generation, and end-to-end replay. Across four leading LLMs, logical plausibility remains near 4/5 while event grounding is only 0.8--2.8/5; return and process quality also disagree. These results expose polished but weakly grounded reasoning that outcome-only evaluation hides. We further argue that P$\rightarrow$E$\rightarrow$R$\rightarrow$D$\rightarrow$O should be a data-system interface, requiring versioned profiles, temporal provenance, inspectable retrieval, decision ledgers, and replayable outcomes. Finance is our stress test for a broader class of personalized, consequential agents.

cs.AI↗
arXiv2026

A Structured Framework for Calibrating Stochastic Car-Following Models: Data Adequacy, Parameter Sensitivity, and Objective Selection

Calibrating a stochastic car-following model is harder than its deterministic counterpart: the loss itself becomes a random variable, so a favorable random realization can be mistaken for a good parameter vector. This paper develops a structured framework for calibrating stochastic car-following models -- a completeness-controlled synthetic design, a corrected variance-based sensitivity analysis (VBSA), and the minimum-realization (MRMIN) calibration protocol -- across two structurally different stochastic mechanisms, QIDM and IDM2D. We test two claims from deterministic calibration -- that a small number of parameters, and the trajectory itself above all, dominates the sensitivity ranking, and that spacing calibration keeps dominating speed calibration once dynamics are stochastic -- and ask whether a model's noise term can be calibrated on its own. In a balanced synthetic experiment, driving-regime completeness has a mean total-effect index on par with the model's most influential parameter and roughly two orders of magnitude above pair identity, extending rather than reversing the deterministic finding on trajectory-identity dominance. Under MRMIN, calibrating only the noise parameter against a population-wide deterministic fit more than doubles median spacing error across 1644 NGSIM trajectories, but fitting the deterministic parameters per trajectory first and calibrating noise on top recovers it. Spacing calibration remains more cross-dimensionally robust than speed calibration on average, but the deterministic guarantee that this dominance can never reverse is violated in 19-26% of trajectories for both mechanisms. A multi-objective screen in relative-error space then favors joint spacing-speed goodness-of-fit functions over single-dimension spacing calibration. Deterministic calibration guarantees should therefore be re-tested, not assumed, once a model is stochastic.

physics.soc-ph↗
arXiv2026

Capek 0.5: An Execution-Centric Vision-Language Model for Embodied Intelligence

Vision-language models are increasingly serving as the reasoning core of embodied agents. Robot execution is inherently iterative: each action reshapes the scene and physical state, continually renewing what must be perceived, reasoned about, and verified. Meeting these demands requires complementary capabilities that differ in supervision signals, prediction formats, and verification criteria. Existing approaches typically develop these capabilities against isolated, task-specific objectives, leaving open how they should be organized and integrated around execution as a whole. We present Capek 0.5, an embodied vision-language model built around an execution-centric capability taxonomy. Rather than organizing training by datasets or tasks, the taxonomy groups embodied capabilities according to their functional roles throughout execution and comprises four capability families: Spatial Reasoning, Temporal Understanding, Action Guidance, and State Verification. Each capability is first acquired by a dedicated specialist through reinforcement learning with verifiable rewards from a shared backbone, and the specialists are then consolidated into a single inference-time model through weight-space merging followed by routed policy-space distillation. We instantiate Capek 0.5 at the 2B and 35B-A3B scales and evaluate it from three complementary perspectives: comprehensive benchmark suites including Capek-StateBench, a new benchmark for state verification; a controlled study of capability retention from specialists to the unified model; and closed-loop evaluation in simulated embodied environments. Capek 0.5 improves the large majority of matched benchmark rows over its initialization, retains all four specialized capabilities in one checkpoint with quantified losses, and transfers to closed-loop embodied task execution.

cs.AI↗
arXiv2026

GeoSAM-Lite: A Lightweight Foundation Model for Onboard Remote Sensing Segmentation

The deployment of large-scale foundation models like Segment Anything Model (SAM) on resource-constrained Earth observation platforms is hindered by prohibitive computational costs and the domain shift between natural and remote sensing imagery. To address these challenges, we propose \textit{Geo}spatial \textit{S}egment \textit{A}nything \textit{M}odel-Lite (GeoSAM-Lite), a lightweight, prompt-free segmentation framework designed for efficient onboard remote sensing segmentation. GeoSAM-Lite incorporates two core innovations: (1) Geospatial-Domain Initialization (Geo-Init), a domain-aware pre-training strategy that distills geospatial priors from a specialized teacher to bridge the domain gap; and (2) Feature Fusion Layers (FFL), which recalibrate spatial features and restore high-frequency boundary cues to overcome the capacity bottlenecks of lightweight backbones. Experiments across representative datasets, with a primary focus on cloud scenarios to evaluate performance under extreme scale variations and complex boundaries, demonstrate that GeoSAM-Lite achieves competitive accuracy while reducing parameters by 92.8\% compared to the heavyweight RSAM-Seg. By establishing a superior Pareto frontier between efficiency and fidelity, GeoSAM-Lite offers a practical solution for real-time segmentation on edge devices.

cs.CV↗
arXiv2026

Green Haven or Risky Venture? Exploring the Connectedness and Hedging of Sustainable Cryptocurrencies and Green Financial Markets

Conventional cryptocurrency often leads to increased energy consumption and carbon emissions, while sustainable cryptocurrencies possess the potential to become a green alternative in portfolio management. This study aims to investigate the time-varying connectedness between sustainable cryptocurrency and green financial markets as well as hedging performance when facing market shocks, including COVID-19 and Russia-Ukraine war. TVP-VAR model with Fourier transform and Multivariate GARCH models are employed. The findings indicate that the pairwise connectedness between the sustainable cryptocurrencies and green financial markets has been at a low level, providing diversification benefits in investment portfolio. Besides, short-term connectedness dominates medium- and long-term connectedness. Sustainable cryptocurrencies show higher hedging effectiveness than traditional cryptocurrency.

stat.AP↗
arXiv2026

Human adaptive variability stabilises collective traffic dynamics

Automated systems are often designed on the assumption that replacing human behavioural variability with precise, uniform algorithmic control improves collective performance. In automotive traffic, this principle underlies commercial adaptive cruise control (ACC). Using two large-scale human-driving experiments comprising 2.95 million car-following observations, a 25-vehicle platoon experiment and a controlled 11-driver protocol, cross-validated with 0.77 million observations from the NGSIM dataset and data from 22 production ACC systems, together with empirically calibrated ACC simulations, we show the opposite: rigid algorithmic uniformity creates systemic fragility. Commercial rule-based controllers amplify small local perturbations into severe stop-and-go waves, increasing fuel consumption and carbon emissions by approximately 2.7- to 5.0-fold across scenarios. Human-driven platoons, by contrast, progressively dissipate disturbances and maintain smoother flow. We identify the behavioural mechanism behind this advantage: human car-following does not follow a fixed proportional spacing rule. Drivers continuously reshape their time-headway distributions across speed regimes, exhibiting a non-monotonic shift from efficiency-oriented to risk-sensitive regulation. This speed-dependent variability generates nonlinear damping that suppresses the synchronisation and propagation of local errors. Our findings challenge the view that human variability is merely suboptimal noise to be eliminated. More broadly, they suggest that robust large-scale interactive AI systems should embed adaptive, human-inspired behavioural flexibility rather than rely on rigid uniformity.

physics.soc-ph↗
arXiv2026

The Moving Target of Urban Equity: Spatiotemporal Demand and Double Disadvantage in Hefei, China

Equitable access to essential urban services is a pillar of modern planning, yet most accessibility models rely strictly on static residential locations, ignoring how demand shifts throughout the daily loop. This study introduces a population-based, temporally differentiated framework to examine the resulting "moving target" of urban equity, focusing on medical facilities and green spaces in Hefei, China. Utilising large-scale mobile phone GPS data, we construct dynamic residential and workplace population exposure surfaces to capture shifting hourly demand. We then evaluate accessibility via network-based travel times paired with a novel per-capita provision metric that accounts for real-time demand competition. We define \textit{double disadvantage} as the co-occurrence of poor spatial accessibility and insufficient per-capita service availability. Counterintuitively, the results reveal that double-disadvantaged areas cluster primarily along the inner suburban belt rather than the remote periphery, where per-capita service provision remains relatively sufficient. Furthermore, temporal shifts drastically alter equity landscapes: daytime workplace concentrations intensely exacerbate demand competition in urban job centres. These findings demonstrate that urban inequality depends heavily on spatiotemporal population flows rather than just the fixed location of services. Ultimately, achieving true urban equity requires dynamic planning interventions that address time-varying demand rather than focusing solely on static, home-based metrics.

physics.soc-ph↗
arXiv2026

DeepInsight: A Unified Evaluation Infrastructure Across the Physical AI Stack

Evaluating a Physical AI stack spans operators that differ by more than three orders of magnitude -- from a single foundation-model decoding step to thousands of physics ticks of whole-body control -- varying orthogonally in modality, reward semantics, and resource profile. No existing framework spans this range, so the stack is evaluated today by stitching together separate harnesses that share neither runtime nor scoring, preserving each segment's local validity but losing the shared identity needed to diagnose cross-layer regressions. We present DeepInsight, an evaluation infrastructure that serves this full spectrum on a single runtime. Rather than homogenize the regimes, it preserves their heterogeneity behind three narrow abstractions -- task, resource, and result -- each realized as one invariant shared by every subsystem: one episode driver, one resource-handle protocol implemented by every expensive backend (LLM inference and sandboxed runtimes alike), and one trace identity scheme under which every event is written. Deployed in production across all three layers of an embodied humanoid stack, this single set of invariants onboards new benchmarks largely by configuration. Where mature peer orchestrators exist -- at the foundation-model end -- it reproduces published references and peer-framework readings within their own spread, runs the same suites faster on a single node, and scales near-linearly across nodes. Its distinctive return is diagnostic: because every layer writes into one shared trace, a regression that begins in one layer and surfaces in another stays localizable on that trace -- a cross-layer payoff no federation of per-segment harnesses can reproduce.

cs.AI↗
arXiv2026

Done, But Not Sure: Disentangling World Completion from Self-Termination in Embodied Agents

Standard embodied evaluations do not independently score whether an agent correctly commits to task completion at episode closure, a capacity we call terminal commitment. Behaviorally distinct failures--never completing the task, completing it but failing to stop, and reporting success without sufficient evidence--collapse into the same benchmark failure. We introduce VIGIL, an evaluation framework that makes terminal commitment independently measurable. Under VIGIL's default protocol, agents observe only egocentric RGB, receive no action-success signals, and must end each episode with a semantic report checked deterministically against hidden world state. This yields two separate scores: world-state completion (W) and benchmark success (B), where B additionally requires a correct terminal report. This decoupling makes four outcome categories distinguishable: missed execution, post-attainment drift, unsupported commitment, and verified success. Across 20 models on 1,000 frozen episodes, systems with comparable W differ by up to 19.7 pp in B: one model converts achieved states into correct reports, while another with near-identical execution drifts past the goal without closing. An action-feedback intervention further tests the separation: execution-oriented signals improve W broadly, yet commitment failures persist in models that do not already ground terminal reports in the achieved state. VIGIL provides a protocol that makes terminal commitment independently visible and scorable.

cs.AI↗
arXiv2025

Twenty-Five Years of the Intelligent Driver Model: Foundations, Extensions, Applications, and Future Directions

The Intelligent Driver Model (IDM), proposed in 2000, has become a foundational tool in traffic flow modeling, renowned for its simplicity, computational efficiency, and ability to capture diverse traffic dynamics. Over the past 25 years, IDM has significantly advanced car-following theory and found extensive application in intelligent transportation systems, including driver assistance systems and autonomous vehicle control. However, IDM's deterministic framework and simplified assumptions face limitations in addressing real-world complexities such as stochastic variability, driver heterogeneity, and mixed traffic conditions. This paper provides a systematic review and critical reflection on IDM's theoretical foundations, academic influence, practical applications, and model extensions. While highlighting IDM's contributions, we emphasize the need to extend the model into a modular and extensible framework. Future directions include integrating stochastic elements, human behavioral insights, and hybrid modeling approaches that combine physics-based structures with data-driven methodologies. By reimagining IDM as a flexible modeling basis, this paper aims to inspire its continued development to meet the demands of intelligent, connected, and increasingly complex traffic systems.

physics.soc-ph↗
arXiv2025

High Finesse Buckled Microcavities

Optical cavities are widely used in modern science and technology to enable a wide range of both quantum and classical applications. Recently, the growing demand for miniaturization and high performance has fueled the exploration of new fabrication methods beyond traditional polishing techniques and macroscopic mirrors. Visible and near-infrared (NIR) wavelengths are particularly important for quantum applications, where achieving low-loss resonators is also more challenging than in the telecom range, presenting unique challenges and opportunities for microscopic cavity systems. Here, we present a novel fabrication method for making NIR microcavities using buckled dielectric membrane mirrors, achieving a record finesse of 0.9 million at 780 nm. We demonstrated flexible device geometries, including singular mirrors and mirror arrays, featuring radii of curvature ranging from 1 mm to 10 mm. The fabrication process offers high uniformity, high yield, and robust performance across a wide range of cavity lengths. Additionally, we can produce easy-to-assemble microcavity packages, with a total volume of ~2 (4) $\,\mathrm{mm}^3\,$, featuring optical modes with a linewidth of 5.16 MHz (570 kHz) and a free spectral range of 3.18 THz (150 GHz). Our results extend the frontier of microcavity fabrication for classical and quantum photonic technologies.

physics.optics↗
arXiv2025

Dynamic-DINO: Fine-Grained Mixture of Experts Tuning for Real-time Open-Vocabulary Object Detection

The Mixture of Experts (MoE) architecture has excelled in Large Vision-Language Models (LVLMs), yet its potential in real-time open-vocabulary object detectors, which also leverage large-scale vision-language datasets but smaller models, remains unexplored. This work investigates this domain, revealing intriguing insights. In the shallow layers, experts tend to cooperate with diverse peers to expand the search space. While in the deeper layers, fixed collaborative structures emerge, where each expert maintains 2-3 fixed partners and distinct expert combinations are specialized in processing specific patterns. Concretely, we propose Dynamic-DINO, which extends Grounding DINO 1.5 Edge from a dense model to a dynamic inference framework via an efficient MoE-Tuning strategy. Additionally, we design a granularity decomposition mechanism to decompose the Feed-Forward Network (FFN) of base model into multiple smaller expert networks, expanding the subnet search space. To prevent performance degradation at the start of fine-tuning, we further propose a pre-trained weight allocation strategy for the experts, coupled with a specific router initialization. During inference, only the input-relevant experts are activated to form a compact subnet. Experiments show that, pretrained with merely 1.56M open-source data, Dynamic-DINO outperforms Grounding DINO 1.5 Edge, pretrained on the private Grounding20M dataset.

cs.CV↗
arXiv2025

RealCam-I2V: Real-World Image-to-Video Generation with Interactive Complex Camera Control

Recent advancements in camera-trajectory-guided image-to-video generation offer higher precision and better support for complex camera control compared to text-based approaches. However, they also introduce significant usability challenges, as users often struggle to provide precise camera parameters when working with arbitrary real-world images without knowledge of their depth nor scene scale. To address these real-world application issues, we propose RealCam-I2V, a novel diffusion-based video generation framework that integrates monocular metric depth estimation to establish 3D scene reconstruction in a preprocessing step. During training, the reconstructed 3D scene enables scaling camera parameters from relative to metric scales, ensuring compatibility and scale consistency across diverse real-world images. In inference, RealCam-I2V offers an intuitive interface where users can precisely draw camera trajectories by dragging within the 3D scene. To further enhance precise camera control and scene consistency, we propose scene-constrained noise shaping, which shapes high-level noise and also allows the framework to maintain dynamic and coherent video generation in lower noise stages. RealCam-I2V achieves significant improvements in controllability and video quality on the RealEstate10K and out-of-domain images. We further enables applications like camera-controlled looping video generation and generative frame interpolation. Project page: https://zgctroy.github.io/RealCam-I2V.

cs.CV↗