arXiv subjects

Zhihao Liu

Publications and source records attributed to Zhihao Liu.

At least 37 records · Page 2Linked to original sources

arXiv2026

Agentic Neuro-Symbolic Planning and Commissioning for Human-in-the-Loop Industrial Robotics with Digital Twins

Flexible robotic automation requires systems that interpret operator intent, verify physical feasibility, and recover from execution failures across both the planning and execution stages. This paper proposes an agentic neuro-symbolic framework for human-in-the-loop industrial robotics, in which LLMs are used for tasks that require language understanding or contextual reasoning, while all verification, sequencing, and execution remain deterministic. The framework adapts the Planner-Generator-Evaluator (PGE) harness pattern from software engineering into a Specifier-Designer-Inspector (SDI) architecture for industrial robotics, combined with LangGraph-based dynamic routing for failure recovery. A two-tier recovery mechanism addresses structure-level replanning through context-aware orchestration and execution-level geometric failures through deterministic recovery skills. A Unity3D digital twin supports human inspection, modification, and re-verification prior to physical execution. Evaluated on natural-language commands across multiple difficulty levels against ten baselines, the proposed method achieves the highest task success. Ablation results confirm that structured command expansion, symbolic verification, selective LLM routing, and recovery skills are each individually necessary.

cs.RO↗

arXiv2026

Aligned but Fragile: Enhancing LLM Safety Robustness via Zeroth-Order Optimization

Safety alignment for large language models (LLMs) aims to reduce harmful or unsafe behavior while preserving general utility. However, recent findings reveal that alignment effects can be fragile: lightweight post-alignment manipulations, such as parameter noise, activation noise, or quantization, can easily weaken the intended safety behavior. Prior efforts to improve robustness have primarily focused on data curation, modified alignment objectives, and safety-critical parameter identification, leaving the role of the optimizer itself largely unexplored. In this paper, we are the first to study the robustness of safety alignment from the perspective of the base optimizer. This optimizer-centric view naturally points to zeroth-order optimization, which provides a robustness-oriented signal by evaluating safety alignment under perturbations. Based on this insight, we propose a hybrid framework that first performs standard first-order safety alignment and then applies zeroth-order refinement to improve robustness. Both theoretically and empirically, we show that only a few zeroth-order refinement steps can enhance robustness while preserving safety alignment. We further improve the efficiency of zeroth-order refinement by exploiting its inherent perturbation-based evaluations to estimate layer-wise robustness sensitivity, enabling the refinement process to concentrate updates on robustness-critical layers with modest training overhead.

cs.AI↗

arXiv2026

Toward Generalizable Graph Learning for 3D Engineering AI: Explainable Workflows for CAE Mode Shape Classification and CFD Field Prediction

Automotive engineering development increasingly relies on heterogeneous 3D data, including finite element (FE) models, body-in-white (BiW) representations, CAD geometry, and CFD meshes. At the same time, engineering teams face growing pressure to shorten development cycles, improve performance and accelerate innovation. Although artificial intelligence (AI) is increasingly explored in this domain, many current methods remain task-specific, difficult to interpret, and hard to reuse across development stages. This paper presents a practical graph learning framework for 3D engineering AI, in which heterogeneous engineering assets are converted into physics-aware graph representations and processed by Graph Neural Networks (GNNs). The framework is designed to support both classification and prediction tasks. The framework is validated on two automotive applications: CAE vibration mode shape classification and CFD aerodynamic field prediction. For CAE vibration mode classification, a region-aware BiW graph supports explainable mode classification across vehicle and FE variants under label scarcity. For CFD aerodynamic field prediction, a physics-informed surrogate predicts pressure and wall shear stress (WSS) across aerodynamic body shape variants, while symmetry preserving down sampling retains accuracy with lower computational cost. The framework also outlines data generation guidance that can help engineers identify which additional simulations or labels are valuable to collect next. These results demonstrate a practical and reusable engineering AI workflow for more trustworthy CAE and CFD decision support.

eess.SY↗

arXiv2026

Automotive Engineering-Centric Agentic AI Workflow Framework

Engineering workflows such as design optimization, simulation-based diagnosis, control tuning, and model-based systems engineering (MBSE) are iterative, constraint-driven, and shaped by prior decisions. Yet many AI methods still treat these activities as isolated tasks rather than as parts of a broader workflow. This paper presents Agentic Engineering Intelligence (AEI), an industrial vision framework that models engineering workflows as constrained, history-aware sequential decision processes in which AI agents support engineer-supervised interventions over engineering toolchains. AEI links an offline phase for engineering data processing and workflow-memory construction with an online phase for workflow-state estimation, retrieval, and decision support. A control-theoretic interpretation is also possible, in which engineering objectives act as reference signals, agents act as workflow controllers, and toolchains provide feedback for intervention selection. Representative automotive use cases in suspension design, reinforcement learning tuning, multimodal engineering knowledge reuse, aerodynamic exploration, and MBSE show how diverse workflows can be expressed within a common formulation. Overall, the paper positions engineering AI as a problem of process-level intelligence and outlines a practical roadmap for future empirical validation in industrial settings.

cs.AI↗

arXiv2026

Shadow in the Cache: Unveiling and Mitigating Privacy Risks of KV-cache in LLM Inference

The Key-Value (KV) cache, which stores intermediate attention computations (Key and Value pairs) to avoid redundant calculations, is a fundamental mechanism for accelerating Large Language Model (LLM) inference. However, this efficiency optimization introduces significant yet underexplored privacy risks. This paper provides the first comprehensive analysis of these vulnerabilities, demonstrating that an attacker can reconstruct sensitive user inputs directly from the KV-cache. We design and implement three distinct attack vectors: a direct Inversion Attack, a more broadly applicable and potent Collision Attack, and a semantic-based Injection Attack. These methods demonstrate the practicality and severity of KV-cache privacy leakage issues. To mitigate this, we propose KV-Cloak, a novel, lightweight, and efficient defense mechanism. KV-Cloak uses a reversible matrix-based obfuscation scheme, combined with operator fusion, to secure the KV-cache. Our extensive experiments show that KV-Cloak effectively thwarts all proposed attacks, reducing reconstruction quality to random noise. Crucially, it achieves this robust security with virtually no degradation in model accuracy and minimal performance overhead, offering a practical solution for trustworthy LLM deployment.

cs.CR↗

arXiv2026

$π_\texttt{RL}$: Online RL Fine-tuning for Flow-based Vision-Language-Action Models

Vision-Language-Action (VLA) models enable robots to understand and perform complex tasks from multimodal input. Although recent work explores using reinforcement learning (RL) to automate the laborious data collection process in scaling supervised fine-tuning (SFT), applying RL to large-scale flow-based VLAs (\eg, $π_0$, $π_{0.5}$) remains challenging due to intractable action log-likelihoods raised from flow matching. We address this challenge with $π_{\texttt{RL}}$, featuring two technical approaches: (1) \textbf{Flow-Noise} models the denoising process as a discrete-time MDP with a learnable noise network for exact log-likelihood computation. (2) \textbf{Flow-SDE} integrates denoising with agent-environment interaction, formulating a two-layer MDP that employs ODE-to-SDE conversion for efficient RL exploration. We evaluate $π_{\texttt{RL}}$ across various benchmarks, with experiments demonstrating that RL yields significant performance improvements in both in-distribution and out-of-distribution settings.

cs.LG↗

arXiv2025

RLinf: Flexible and Efficient Large-scale Reinforcement Learning via Macro-to-Micro Flow Transformation

Reinforcement learning (RL) has demonstrated immense potential in advancing artificial general intelligence, agentic intelligence, and embodied intelligence. However, the inherent heterogeneity and dynamicity of RL workflows often lead to low hardware utilization and slow training on existing systems. In this paper, we present RLinf, a high-performance RL training system based on our key observation that the major roadblock to efficient RL training lies in system flexibility. To maximize flexibility and efficiency, RLinf is built atop a novel RL system design paradigm called macro-to-micro flow transformation (M2Flow), which automatically breaks down high-level, easy-to-compose RL workflows at both the temporal and spatial dimensions, and recomposes them into optimized execution flows. Supported by RLinf worker's adaptive communication capability, we devise context switching and elastic pipelining to realize M2Flow transformation, and a profiling-guided scheduling policy to generate optimal execution plans. Extensive evaluations on both reasoning RL and embodied RL tasks demonstrate that RLinf consistently outperforms state-of-the-art systems, achieving $1.07\times-2.43\times$ speedup in end-to-end training throughput.

cs.LG↗

arXiv2025

FlareX: A Physics-Informed Dataset for Lens Flare Removal via 2D Synthesis and 3D Rendering

Lens flare occurs when shooting towards strong light sources, significantly degrading the visual quality of images. Due to the difficulty in capturing flare-corrupted and flare-free image pairs in the real world, existing datasets are typically synthesized in 2D by overlaying artificial flare templates onto background images. However, the lack of flare diversity in templates and the neglect of physical principles in the synthesis process hinder models trained on these datasets from generalizing well to real-world scenarios. To address these challenges, we propose a new physics-informed method for flare data generation, which consists of three stages: parameterized template creation, the laws of illumination-aware 2D synthesis, and physical engine-based 3D rendering, which finally gives us a mixed flare dataset that incorporates both 2D and 3D perspectives, namely FlareX. This dataset offers 9,500 2D templates derived from 95 flare patterns and 3,000 flare image pairs rendered from 60 3D scenes. Furthermore, we design a masking approach to obtain real-world flare-free images from their corrupted counterparts to measure the performance of the model on real-world images. Extensive experiments demonstrate the effectiveness of our method and dataset.

cs.CV↗

arXiv2025

A Unified Framework for Optimizing Uniformly Controlled Structures in Quantum Circuits

Quantum unitaries of the form $Σ_{c}\ket{c}\bra{c}\otimes U_{c}$ are ubiquitous in quantum algorithms. This class encompasses not only standard uniformly controlled gates (UCGs) but also a wide range of circuits with uniformly controlled structures. However, their circuit-depth and gate-count complexities have not been systematically analyzed within a unified framework. In this work, we study the general decomposition problem for UCG and UCG-like structure. We then introduce the restricted Uniformly Controlled Gates (rUCGs) as a unified algebraic model, defined by a 2-divisible Abelian group that models the controlled gate set. This model captures uniformly controlled rotations, multi-qubit uniformly controlled gates, and diagonal unitaries. Furthermore, this model also naturally incorporates k-sparse version (k-rUCGs), where only a subset of control qubits participate in each multi-qubit gate. Building on this algebraic model, we develop a general framework. For an n-control rUCG, the framework reduce the gate complexity from ${O(n2^n)}$ to ${O(2^n})$ and the circuit depth from ${O(2^n\log n)}$ to ${O(2^n\log n/n)}$. The framework further provides systematic size and depth bounds for k-rUCGs by exploiting sparsity in the control space, with same optimization coefficient as rUCG, respectively. Empirical evaluations on representative QAOA circuits confirm reductions in depth and size, which highlight that the rUCG model and its associated decomposition framework unify circuits previously considered structurally distinct under a single, asymptotically optimal synthesis paradigm.

quant-ph↗

arXiv2025

Quasi-linear magnetoresistance and paramagnetic singularity in Hypervalent Bismuthide

Materials featuring hypervalent bismuth motifs have generated immense interest due to their extraordinary electronic structure and exotic quantum transport. In this study, we synthesized high-quality single crystals of La3ScBi5 characterized by one-dimensional hypervalent bismuth chains and performed a systematic investigation of the magnetoresistive behavior and quantum oscillations. The metallic La3ScBi5 exhibits a low-temperature plateau of electrical resistivity and quasi-linear positive magnetoresistance, with anisotropic magnetoresistive behaviors suggesting the presence of anisotropic Fermi surfaces. This distinctive transport phenomenon is perfectly elucidated by first-principles calculations utilizing the semiclassical Boltzmann transport theory. Furthermore, the nonlinear Hall resistivity pointed towards a multiband electronic structure, characterized by the coexistence of electron and hole carriers, which is further supported by our first-principles calculations. Angle-dependent de Haas-van Alphen oscillations are crucial for further elucidating its Fermiology and topological characteristics. Intriguingly, magnetization measurements unveiled a notable paramagnetic singularity at low fields, which might suggest the nontrivial nature of the surface states. Our findings underscore the interplay between transport phenomena and the unique electronic structure of hypervalent bismuthide La3ScBi5, opening avenues for exploring novel electronic applications.

cond-mat.mtrl-sci↗

arXiv2025

DisasterM3: A Remote Sensing Vision-Language Dataset for Disaster Damage Assessment and Response

Large vision-language models (VLMs) have made great achievements in Earth vision. However, complex disaster scenes with diverse disaster types, geographic regions, and satellite sensors have posed new challenges for VLM applications. To fill this gap, we curate a remote sensing vision-language dataset (DisasterM3) for global-scale disaster assessment and response. DisasterM3 includes 26,988 bi-temporal satellite images and 123k instruction pairs across 5 continents, with three characteristics: 1) Multi-hazard: DisasterM3 involves 36 historical disaster events with significant impacts, which are categorized into 10 common natural and man-made disasters. 2)Multi-sensor: Extreme weather during disasters often hinders optical sensor imaging, making it necessary to combine Synthetic Aperture Radar (SAR) imagery for post-disaster scenes. 3) Multi-task: Based on real-world scenarios, DisasterM3 includes 9 disaster-related visual perception and reasoning tasks, harnessing the full potential of VLM's reasoning ability with progressing from disaster-bearing body recognition to structural damage assessment and object relational reasoning, culminating in the generation of long-form disaster reports. We extensively evaluated 14 generic and remote sensing VLMs on our benchmark, revealing that state-of-the-art models struggle with the disaster tasks, largely due to the lack of a disaster-specific corpus, cross-sensor gap, and damage object counting insensitivity. Focusing on these issues, we fine-tune four VLMs using our dataset and achieve stable improvements across all tasks, with robust cross-sensor and cross-disaster generalization capabilities. The code and data are available at: https://github.com/Junjue-Wang/DisasterM3.

cs.CV↗

arXiv2025

BurstDeflicker: A Benchmark Dataset for Flicker Removal in Dynamic Scenes

Flicker artifacts in short-exposure images are caused by the interplay between the row-wise exposure mechanism of rolling shutter cameras and the temporal intensity variations of alternating current (AC)-powered lighting. These artifacts typically appear as uneven brightness distribution across the image, forming noticeable dark bands. Beyond compromising image quality, this structured noise also affects high-level tasks, such as object detection and tracking, where reliable lighting is crucial. Despite the prevalence of flicker, the lack of a large-scale, realistic dataset has been a significant barrier to advancing research in flicker removal. To address this issue, we present BurstDeflicker, a scalable benchmark constructed using three complementary data acquisition strategies. First, we develop a Retinex-based synthesis pipeline that redefines the goal of flicker removal and enables controllable manipulation of key flicker-related attributes (e.g., intensity, area, and frequency), thereby facilitating the generation of diverse flicker patterns. Second, we capture 4,000 real-world flicker images from different scenes, which help the model better understand the spatial and temporal characteristics of real flicker artifacts and generalize more effectively to wild scenarios. Finally, due to the non-repeatable nature of dynamic scenes, we propose a green-screen method to incorporate motion into image pairs while preserving real flicker degradation. Comprehensive experiments demonstrate the effectiveness of our dataset and its potential to advance research in flicker removal.

cs.CV↗

arXiv2025

Complex field-, temperature-, and angle-dependent Hall effects from intrinsic Fermi surface revealed by first-principles calculations

The Hall effect, ever intriguing since its discovery, has spurred the exploration of its phenomena, intensified by advances in topology and novel materials. Differentiating the ordinary Hall effect from extraordinary properties like the anomalous Hall effect (AHE) is challenging, especially in materials with topological origins. In our study, we leverage semiclassical Boltzmann transport theory and first-principles calculations within the relaxation time approximation to analyze Hall effects comprehensively. We have found that the complex magnetic field dependence of ordinary Hall effect, including the sign reversals, appearing of plateau and nonlinearity, can be understood and reproduced by our approach both for multiband models and realistic topological materials of ZrSiS and PtTe2. The Hall resistivity versus temperature and magnetic fields can be well scaled, similar to Kohler's rule for longitudinal resistivity. This methodology can also accurately model the angular dependent Hall effects such as planar Hall effects of bismuth. These findings indicate that the dependencies of various Hall effects and magnetoresistance on magnetic fields are mainly determined by the details of Fermi surface and the relaxation time. The intrinsic Fermi surface determines the carriers' density, type, and velocity, while the later is mostly influenced by extrinsic factors, such as quality of sample with defects, impurities, and domains. This insight might simplify the understanding of several seemingly complex transport phenomena in nonmagnetic materials, with no need for hypotheses of other sophisticated mechanisms, such as magnetization-induced AHE, Lifshitz transition-induced changes in carrier type, exotic orders like charge density wave, or some delicate scattering of carriers with chiral or nonreciprocal dependence. Finally, we also discussed the Hall effects contribute from the Berry curvature.

cond-mat.mtrl-sci↗

arXiv2025

A combined First-principles and Boltzmann transport theory methodology for studying magnetotransport in magnetic materials

Unusual magnetotransport behaviors, such as temperature-dependent negative magnetoresistance (MR) and bowtie-shaped MR, have puzzled us for a long time. Although several mechanisms have been proposed to explain these phenomena, the absence of comprehensive quantitative calculations has made these explanations less convincing. In our work, we introduce a methodology to study magnetotransport behaviors in magnetic materials. This approach integrates the anomalous Hall conductivity induced by the Berry curvature, with a multi-band ordinary conductivity tensor, employing a combination of first-principles calculations and semi-classical Boltzmann transport theory. Our method also incorporates the temperature dependence of the relaxation time and the anomalous Hall conductivity, as well as the field dependence of the anomalous Hall conductivity. We initially test this approach on models, obtaining distinct behaviors of magnetoresistance and Hall resistivity across several types of magnetic materials, and then apply it to a Weyl semimetal, Co3Sn2S2. The results, which align well with experimental observations in terms of magnetic field and temperature dependencies, demonstrate the efficacy of our approach. This methodology provides a comprehensive and efficient means to understand the underlying mechanisms of the unusual complex behaviors observed in magnetotransport measurements in magnetic materials.

cond-mat.mtrl-sci↗

arXiv2025

AIM 2025 Challenge on High FPS Motion Deblurring: Methods and Results

This paper presents a comprehensive review of the AIM 2025 High FPS Non-Uniform Motion Deblurring Challenge, highlighting the proposed solutions and final results. The objective of this challenge is to identify effective networks capable of producing clearer and visually compelling images in diverse and challenging conditions, by learning representative visual cues for complex aggregations of motion types. A total of 68 participants registered for the competition, and 9 teams ultimately submitted valid entries. This paper thoroughly evaluates the state-of-the-art advances in high-FPS single image motion deblurring, showcasing the significant progress in the field, while leveraging samples of the novel dataset, MIORe, that introduces challenging examples of movement patterns.

cs.CV↗

arXiv2025

Module-Aware Parameter-Efficient Machine Unlearning on Transformers

Transformer has become fundamental to a vast series of pre-trained large models that have achieved remarkable success across diverse applications. Machine unlearning, which focuses on efficiently removing specific data influences to comply with privacy regulations, shows promise in restricting updates to influence-critical parameters. However, existing parameter-efficient unlearning methods are largely devised in a module-oblivious manner, which tends to inaccurately identify these parameters and leads to inferior unlearning performance for Transformers. In this paper, we propose {\tt MAPE-Unlearn}, a module-aware parameter-efficient machine unlearning approach that uses a learnable pair of masks to pinpoint influence-critical parameters in the heads and filters of Transformers. The learning objective of these masks is derived by desiderata of unlearning and optimized through an efficient algorithm featured by a greedy search with a warm start. Extensive experiments on various Transformer models and datasets demonstrate the effectiveness and robustness of {\tt MAPE-Unlearn} for unlearning.

cs.LG↗

arXiv2025

Towards Evaluation for Real-World LLM Unlearning

This paper analyzes the limitations of existing unlearning evaluation metrics in terms of practicality, exactness, and robustness in real-world LLM unlearning scenarios. To overcome these limitations, we propose a new metric called Distribution Correction-based Unlearning Evaluation (DCUE). It identifies core tokens and corrects distributional biases in their confidence scores using a validation set. The evaluation results are quantified using the Kolmogorov-Smirnov test. Experimental results demonstrate that DCUE overcomes the limitations of existing metrics, which also guides the design of more practical and reliable unlearning algorithms in the future.

cs.AI↗

arXiv2025

Retrieving snow depth distribution by downscaling ERA5 Reanalysis with ICESat-2 laser altimetry

Estimating the variability of seasonal snow cover, in particular snow depth in remote areas, poses significant challenges due to limited spatial and temporal data availability. This study uses snow depth measurements from the ICESat-2 satellite laser altimeter, which are sparse in both space and time, and incorporates them with climate reanalysis data into a downscaling-calibration scheme to produce monthly gridded snow depth maps at microscale (10 m). Snow surface elevation measurements from ICESat-2 along profiles are compared to a digital elevation model to determine snow depth at each point. To efficiently turn sparse measurements into snow depth maps, a regression model is fitted to establish a relationship between the retrieved snow depth and the corresponding ERA5 Land snow depth. This relationship, referred to as subgrid variability, is then applied to downscale the monthly ERA5 Land snow depth data. The method can provide timeseries of monthly snow depth maps for the entire ERA5 time range (since 1950). The validation of downscaled snow depth data was performed at an intermediate scale (100 m x 500 m) using datasets from airborne laser scanning (ALS) in the Hardangervidda region of southern Norway. Results show that snow depth prediction achieved R2 values ranging from 0.74 to 0.88 (post-calibration). The method relies on globally available data and is applicable to other snow regions above the treeline. Though requiring area-specific calibration, our approach has the potential to provide snow depth maps in areas where no such data exist and can be used to extrapolate existing snow surveys in time and over larger areas. With this, it can offer valuable input data for hydrological, ecological or permafrost modeling tasks.

physics.geo-ph↗