arXiv subjects

Shihao Li

Publications and source records attributed to Shihao Li.

At least 55 records · Page 3Linked to original sources

arXiv2026

An LLM-Assisted Multi-Agent Control Framework for Roll-to-Roll Manufacturing Systems

Roll-to-roll manufacturing requires precise tension and velocity control to ensure product quality, yet controller commissioning and adaptation remain time-intensive processes dependent on expert knowledge. This paper presents an LLM-assisted multi-agent framework that automates control system design and adaptation for R2R systems while maintaining safety. The framework operates through five phases: system identification from operational data, automated controller selection and tuning, sim-to-real adaptation with safety verification, continuous monitoring with diagnostic capabilities, and periodic model refinement. Experimental validation on a R2R system demonstrates successful tension regulation and velocity tracking under significant model uncertainty, with the framework achieving performance convergence through iterative adaptation. The approach reduces manual tuning effort while providing transparent diagnostic information for maintenance planning, offering a practical pathway for integrating AI-assisted automation in manufacturing control systems.

eess.SY↗

arXiv2026

Optical manipulation of valley coherence via Landau level transitions in black phosphorus and WTe2 monolayers

Valley coherence is of great significance for exploring fundamental quantum phenomena and developing next-generation valleytronic devices. Herein, we theoretically investigate the valley quantum interference engineered by inter-Landau level (LL) transitions in black phosphorus (BP) and WTe2 monolayers. In contrast to the non-Landau-quantized regime, valley quantum interference is enhanced by over 20-fold, or even significantly stronger, in virtue of striking anisotropic environment. Such anisotropy originates from the distinct electron transition probabilities along the armchair and zigzag directions of BP and WTe2 monolayers. Especially, BP is capable of more effectively strengthening the valley quantum interference response due to its greater directional disparity in electron transition probabilities. The interference fringes also present distinct spectral profiles (e.g., different dip and peak numbers in one interference period) owing to different transition selection rules in BP and WTe2 monolayers. In spite of these discrepancies, normalized interference intensities follow two exponential functions of magnetic field and Landau level index for all the transitions δn = n'- n = -4, -2, 0, +2, +4 (where n and n' indicate the LL indexes of valence and conduction bands, respectively), and the interference spectra exhibit C2 rotational symmetry about the crystallographic azimuthal angle of 90°.

cond-mat.mes-hall↗

arXiv2026

SCAFFOLD-CEGIS: Preventing Latent Security Degradation in LLM-Driven Iterative Code Refinement

The application of large language models to code generation has evolved from one-shot generation to iterative refinement, yet the evolution of security throughout iteration remains insufficiently understood. Through comparative experiments on three mainstream LLMs, this paper reveals the iterative refinement paradox: specification drift during multi-objective optimization causes security to degrade gradually over successive iterations. Taking GPT-4o as an example, 43.7 % of iteration chains contain more vulnerabilities than the baseline after ten rounds, and cross-model experiments show that this phenomenon is prevalent. Further analysis shows that simply introducing static application security testing (SAST) gating cannot effectively suppress degradation; instead, it increases the latent security degradation rate from 12.5% under the unprotected baseline to 20.8 %. The root cause is that static-analysis rules cannot cover structural degradations such as the removal of defensive logic or the weakening of exception handling. To address this problem, we propose the SCAFFOLD-CEGIS framework. Drawing on the counterexample-guided inductive synthesis (CEGIS) paradigm, the framework adopts a multi-agent collaborative architecture that transforms security constraints from implicit prompts into explicit verifiable constraints. It automatically identifies and solidifies security-critical elements as hard constraints through semantic anchoring, enforces safety monotonicity through four-layer gated verification, and continuously assimilates experience from failures. Comparative experiments against six existing defense methods show that the full framework reduces the latent security degradation rate to 2.1% and achieves a safety monotonicity rate of 100%.

cs.CR↗

arXiv2026

AnyPcc: Compressing Any Point Cloud with a Single Universal Model

Generalization remains a critical challenge in deep learning-based point cloud geometry compression. While existing methods perform well on standard benchmarks, their performance collapses in real-world scenarios due to two fundamental limitations: the lack of context models that are robust across diverse data densities, and the inability to efficiently adapt to out-of-distribution (OOD) data. To overcome both challenges, we introduce AnyPcc, a universal point cloud compression framework. AnyPcc first employs a Universal Context Model that leverages coarse-grained spatial priors with fine-grained channel priors to ensure robust context modeling across the entire density spectrum. Second, our novel Instance-Adaptive Fine-Tuning (IAFT) strategy tackles OOD data by synergizing explicit and implicit compression paradigms. For each instance, it fine-tunes a small subset of network weights and transmits them within the bitstream. The minimal bitrate overhead from these weights is significantly outweighed by the resulting gains in geometry compression. Extensive experiments on a benchmark of 15 diverse datasets confirm that AnyPcc sets a new state-of-the-art in point cloud compression while maintaining low complexity. Our code and datasets have been released to encourage reproducible research.

cs.CV↗

arXiv2026

OmniVideoBench: Towards Audio-Visual Understanding Evaluation for Omni MLLMs

Recent advances in multimodal large language models (MLLMs) have demonstrated substantial potential in video understanding. However, existing benchmarks fail to comprehensively evaluate synergistic reasoning capabilities across audio and visual modalities, often neglecting either one of the modalities or integrating them in a logically inconsistent manner. To bridge this gap, we introduce OmniVideoBench, a large-scale and rigorously designed benchmark dedicated to assessing synergistic audio-visual understanding, with a strong emphasis on modality complementarity and logical consistency. Specifically, OmniVideoBench comprises 1000 high-quality question-answer(QA) pairs, each annotated with step-by-step reasoning traces, derived from 628 diverse videos ranging from several seconds to 30 minutes, and manually verified to guarantee complete correctness and uniqueness. Moreover, OmniVideoBench encompasses 13 carefully designed question types, covering temporal reasoning, spatial localization, counting, causal inference, summarization, and beyond, thereby capturing the essential challenges of video understanding. Evaluation of multiple MLLMs on OmniVideoBench reveals a pronounced gap between model performance and human reasoning, with open-source models lagging significantly behind their closed-source counterparts, underscoring the inherent difficulty of genuine audio-visual reasoning. We will release OmniVideoBench to foster the development of MLLMs with stronger and more generalizable reasoning capabilities.

cs.AI↗

arXiv2026

Curriculum-Based Soft Actor-Critic for Multi-Section R2R Tension Control

Precise tension control in roll-to-roll (R2R) manufacturing is difficult under varying operating conditions and process uncertainty. This paper presents a curriculum-based Soft Actor-Critic (SAC) controller for multi-section R2R tension control. The policy is trained in three phases with progressively wider reference ranges, from 27 to 33 N to the full operating envelope of 20 to 40 N, so it can generalize across nominal and disturbed conditions. On a three-section R2R benchmark, the learned controller achieves accurate tracking in nominal operation and handles large disturbances, including 20 N to 40 N step changes, with a single policy and no scenario-specific retuning. These results indicate that curriculum-trained SAC is a practical alternative to model-based control when system parameters vary and process uncertainty is significant.

eess.SY↗

arXiv2026

DiffPlace: Street View Generation via Place-Controllable Diffusion Model Enhancing Place Recognition

Generative models have advanced significantly in realistic image synthesis, with diffusion models excelling in quality and stability. Recent multi-view diffusion models improve 3D-aware street view generation, but they struggle to produce place-aware and background-consistent urban scenes from text, BEV maps, and object bounding boxes. This limits their effectiveness in generating realistic samples for place recognition tasks. To address these challenges, we propose DiffPlace, a novel framework that introduces a place-ID controller to enable place-controllable multi-view image generation. The place-ID controller employs linear projection, perceiver transformer, and contrastive learning to map place-ID embeddings into a fixed CLIP space, allowing the model to synthesize images with consistent background buildings while flexibly modifying foreground objects and weather conditions. Extensive experiments, including quantitative comparisons and augmented training evaluations, demonstrate that DiffPlace outperforms existing methods in both generation quality and training support for visual place recognition. Our results highlight the potential of generative models in enhancing scene-level and place-aware synthesis, providing a valuable approach for improving place recognition in autonomous driving

cs.CV↗

arXiv2026

Vibe AIGC: A New Paradigm for Content Generation via Agentic Orchestration

For the past decade, the trajectory of generative artificial intelligence (AI) has been dominated by a model-centric paradigm driven by scaling laws. Despite significant leaps in visual fidelity, this approach has encountered a ``usability ceiling'' manifested as the Intent-Execution Gap (i.e., the fundamental disparity between a creator's high-level intent and the stochastic, black-box nature of current single-shot models). In this paper, inspired by the Vibe Coding, we introduce the \textbf{Vibe AIGC}, a new paradigm for content generation via agentic orchestration, which represents the autonomous synthesis of hierarchical multi-agent workflows. Under this paradigm, the user's role transcends traditional prompt engineering, evolving into a Commander who provides a Vibe, a high-level representation encompassing aesthetic preferences, functional logic, and etc. A centralized Meta-Planner then functions as a system architect, deconstructing this ``Vibe'' into executable, verifiable, and adaptive agentic pipelines. By transitioning from stochastic inference to logical orchestration, Vibe AIGC bridges the gap between human imagination and machine execution. We contend that this shift will redefine the human-AI collaborative economy, transforming AI from a fragile inference engine into a robust system-level engineering partner that democratizes the creation of complex, long-horizon digital assets.

cs.AI↗

arXiv2026

WaveMan: mmWave-Based Room-Scale Human Interaction Perception for Humanoid Robots

Reliable humanoid-robot interaction (HRI) in household environments is constrained by two fundamental requirements, namely robustness to unconstrained user positions and preservation of user privacy. Millimeter-wave (mmWave) sensing inherently supports privacy-preserving interaction, making it a promising modality for room-scale HRI. However, existing mmWave-based interaction-sensing systems exhibit poor spatial generalization at unseen distances or viewpoints. To address this challenge, we introduce WaveMan, a spatially adaptive room-scale perception system that restores reliable human interaction sensing across arbitrary user positions. WaveMan integrates viewpoint alignment and spectrogram enhancement for spatial consistency, with dual-channel attention for robust feature extraction. Experiments across five participants show that, under fixed-position evaluation, WaveMan achieves the same cross-position accuracy as the baseline with five times fewer training positions. In random free-position testing, accuracy increases from 33.00% to 94.33%, enabled by the proposed method. These results demonstrate the feasibility of reliable, privacy-preserving interaction for household humanoid robots across unconstrained user positions.

cs.RO↗

arXiv2026

MT-Video-Bench: A Holistic Video Understanding Benchmark for Evaluating Multimodal LLMs in Multi-Turn Dialogues

The recent development of Multimodal Large Language Models (MLLMs) has significantly advanced AI's ability to understand visual modalities. However, existing evaluation benchmarks remain limited to single-turn question answering, overlooking the complexity of multi-turn dialogues in real-world scenarios. To bridge this gap, we introduce MT-Video-Bench, a holistic video understanding benchmark for evaluating MLLMs in multi-turn dialogues. Specifically, our MT-Video-Bench mainly assesses 6 core competencies that focus on perceptivity and interactivity, encompassing 1,000 meticulously curated multi-turn dialogues from diverse domains. These capabilities are rigorously aligned with real-world applications, such as interactive sports analysis and multi-turn video-based intelligent tutoring. With MT-Video-Bench, we extensively evaluate various state-of-the-art open-source and closed-source MLLMs, revealing their significant performance discrepancies and limitations in handling multi-turn video dialogues. The benchmark will be publicly available to foster future research.

cs.CV↗

arXiv2026

DCG ReID: Disentangling Collaboration and Guidance Fusion Representations for Multi-modal Vehicle Re-Identification

Multi-modal vehicle Re-Identification (ReID) aims to leverage complementary information from RGB, Near Infrared (NIR), and Thermal Infrared (TIR) modalities to retrieve the same vehicle. The challenges of multi-modal vehicle ReID arise from the uncertainty of modality quality distribution induced by inherent discrepancies across modalities, resulting in distinct conflicting fusion requirements for data with balanced and unbalanced quality distributions. Existing methods handle all multi-modal data within a single fusion model, overlooking the different needs of the two data types and making it difficult to decouple the conflict between intra-class consistency and inter-modal heterogeneity. To this end, we propose Disentangle Collaboration and Guidance Fusion Representations for Multi-modal Vehicle ReID (DCG-ReID). Specifically, to disentangle heterogeneous quality-distributed modal data without mutual interference, we first design the Dynamic Confidence-based Disentangling Weighting (DCDW) mechanism: dynamically reweighting three-modal contributions via interaction-derived modal confidence to build a disentangled fusion framework. Building on DCDW, we develop two scenario-specific fusion strategies: (1) for balanced quality distributions, Collaboration Fusion Module (CFM) mines pairwise consensus features to capture shared discriminative information and boost intra-class consistency; (2) for unbalanced distributions, Guidance Fusion Module (GFM) implements differential amplification of modal discriminative disparities to reinforce dominant modality advantages, guide auxiliary modalities to mine complementary discriminative info, and mitigate inter-modal divergence to boost multi-modal joint decision performance. Extensive experiments on three multi-modal ReID benchmarks (WMVeID863, MSVR310, RGBNT100) validate the effectiveness of our method. Code will be released upon acceptance.

cs.CV↗

arXiv2025

OmniBench: Towards The Future of Universal Omni-Language Models

Recent advancements in multimodal large language models (MLLMs) have aimed to integrate and interpret data across diverse modalities. However, the capacity of these models to concurrently process and reason about multiple modalities remains underexplored, partly due to the lack of comprehensive modality-wise benchmarks. We introduce OmniBench, a novel benchmark designed to rigorously evaluate models' ability to recognize, interpret, and reason across visual, acoustic, and textual inputs simultaneously. We define language models capable of such tri-modal processing as the omni-language models (OLMs). OmniBench is distinguished by high-quality human annotations, ensuring that accurate responses require integrated understanding and reasoning across all three modalities. Our main findings reveal that: i) open-source OLMs exhibit critical limitations in instruction-following and reasoning capabilities within tri-modal contexts; and ii) most baselines models perform poorly (below 50% accuracy) even when provided with alternative textual representations of images or/and audio. These results suggest that the ability to construct a consistent context from text, image, and audio is often overlooked in existing MLLM training paradigms. To address this gap, we curate an instruction tuning dataset of 84.5K training samples, OmniInstruct, for training OLMs to adapt to tri-modal contexts. We advocate for future research to focus on developing more robust tri-modal integration techniques and training strategies to enhance OLMs. Codes, data and live leaderboard could be found at https://m-a-p.ai/OmniBench.

cs.CL↗

arXiv2025

Adaptive Trajectory Bundle Method for Roll-to-Roll Manufacturing Systems

Roll-to-roll (R2R) manufacturing requires precise tension and velocity control under operational constraints. Model predictive control demands gradient computation, while sampling-based methods like MPPI struggle with hard constraint satisfaction. This paper presents an adaptive trajectory bundle method that achieves rigorous constraint handling through derivative-free sequential convex programming. The approach approximates nonlinear dynamics and costs via interpolated sample bundles, replacing Taylor-series linearization with function-value interpolation. Adaptive trust region and penalty mechanisms automatically adjust based on constraint violation metrics, eliminating manual tuning. We establish convergence guarantees proving finite-time feasibility and convergence to stationary points of the constrained problem. Simulations on a six-zone R2R system demonstrate that the adaptive method achieves 4.3\% lower tension RMSE than gradient-based MPC and 11.1\% improvement over baseline TBM in velocity transients, with superior constraint satisfaction compared to MPPI variants. Experimental validation on an R2R dry transfer system confirms faster settling and reduced overshoot relative to LQR and non-adaptive TBM.

eess.SY↗

arXiv2025

Natural Geometry of Robust Data Attribution: From Convex Models to Deep Networks

Data attribution methods identify which training examples are responsible for a model's predictions, but their sensitivity to distributional perturbations undermines practical reliability. We present a unified framework for certified robust attribution that extends from convex models to deep networks. For convex settings, we derive Wasserstein-Robust Influence Functions (W-RIF) with provable coverage guarantees. For deep networks, we demonstrate that Euclidean certification is rendered vacuous by spectral amplification -- a mechanism where the inherent ill-conditioning of deep representations inflates Lipschitz bounds by over $10{,}000\times$. This explains why standard TRAK scores, while accurate point estimates, are geometrically fragile: naive Euclidean robustness analysis yields 0\% certification. Our key contribution is the Natural Wasserstein metric, which measures perturbations in the geometry induced by the model's own feature covariance. This eliminates spectral amplification, reducing worst-case sensitivity by $76\times$ and stabilizing attribution estimates. On CIFAR-10 with ResNet-18, Natural W-TRAK certifies 68.7\% of ranking pairs compared to 0\% for Euclidean baselines -- to our knowledge, the first non-vacuous certified bounds for neural network attribution. Furthermore, we prove that the Self-Influence term arising from our analysis equals the Lipschitz constant governing attribution stability, providing theoretical grounding for leverage-based anomaly detection. Empirically, Self-Influence achieves 0.970 AUROC for label noise detection, identifying 94.1\% of corrupted labels by examining just the top 20\% of training data.

cs.LG↗

arXiv2025

MVU-Eval: Towards Multi-Video Understanding Evaluation for Multimodal LLMs

The advent of Multimodal Large Language Models (MLLMs) has expanded AI capabilities to visual modalities, yet existing evaluation benchmarks remain limited to single-video understanding, overlooking the critical need for multi-video understanding in real-world scenarios (e.g., sports analytics and autonomous driving). To address this significant gap, we introduce MVU-Eval, the first comprehensive benchmark for evaluating Multi-Video Understanding for MLLMs. Specifically, our MVU-Eval mainly assesses eight core competencies through 1,824 meticulously curated question-answer pairs spanning 4,959 videos from diverse domains, addressing both fundamental perception tasks and high-order reasoning tasks. These capabilities are rigorously aligned with real-world applications such as multi-sensor synthesis in autonomous systems and cross-angle sports analytics. Through extensive evaluation of state-of-the-art open-source and closed-source models, we reveal significant performance discrepancies and limitations in current MLLMs' ability to perform understanding across multiple videos. The benchmark will be made publicly available to foster future research.

cs.CV↗

arXiv2025

FabasedVC: Enhancing Voice Conversion with Text Modality Fusion and Phoneme-Level SSL Features

In voice conversion (VC), it is crucial to preserve complete semantic information while accurately modeling the target speaker's timbre and prosody. This paper proposes FabasedVC to achieve VC with enhanced similarity in timbre, prosody, and duration to the target speaker, as well as improved content integrity. It is an end-to-end VITS-based VC system that integrates relevant textual modality information, phoneme-level self-supervised learning (SSL) features, and a duration predictor. Specifically, we employ a text feature encoder to encode attributes such as text, phonemes, tones and BERT features. We then process the frame-level SSL features into phoneme-level features using two methods: average pooling and attention mechanism based on each phoneme's duration. Moreover, a duration predictor is incorporated to better align the speech rate and prosody of the target speaker. Experimental results demonstrate that our method outperforms competing systems in terms of naturalness, similarity, and content integrity.

cs.SD↗

arXiv2025

Algorithm-Relative Trajectory Valuation in Policy Gradient Control

We study how trajectory value depends on the learning algorithm in policy-gradient control. Using Trajectory Shapley in an uncertain LQR, we find a negative correlation between Persistence of Excitation (PE) and marginal value under vanilla REINFORCE ($r\approx-0.38$). We prove a variance-mediated mechanism: (i) for fixed energy, higher PE yields lower gradient variance; (ii) near saddles, higher variance increases escape probability, raising marginal contribution. When stabilized (state whitening or Fisher preconditioning), this variance channel is neutralized and information content dominates, flipping the correlation positive ($r\approx+0.29$). Hence, trajectory value is algorithm-relative. Experiments validate the mechanism and show decision-aligned scores (Leave-One-Out) complement Shapley for pruning, while Shapley identifies toxic subsets.

cs.LG↗

arXiv2025

From Natural Language to Certified H-infinity Controllers: Integrating LLM Agents with LMI-Based Synthesis

We present \textsc{S2C} (Specification-to-Certified-Controller), a multi-agent framework that maps natural-language requirements to certified $\mathcal{H}_\infty$ state-feedback controllers via LMI synthesis. \textsc{S2C} coordinates five roles -- \textit{SpecInt} (spec extraction), \textit{Solv} (bounded-real lemma (BRL) LMI), \textit{Tester} (Monte Carlo and frequency-domain checks), \textit{Adapt} (spec refinement), and \textit{CodeGen} (deployable code). The loop is stabilized by a severity- and iteration-aware $γ$-floor guardrail and a decay-rate region constraint enforcing $\Reλ(A{+}BK)<-α$ with $α=3.9/T_s$ derived from settling-time targets. For state feedback, verification reports disturbance rejection $\big\|C\,(sI-(A{+}BK))^{-1}E\big\|_\infty$ alongside time-domain statistics; discrete benchmarks are converted to continuous time via a Tustin (bilinear) transform when needed. On 14 COMPleib problems, \textsc{S2C} attains \textbf{100\%} synthesis success and \textbf{100\%} convergence within six iterations, with strong decay-rate satisfaction and near-target certified $\mathcal{H}_\infty$ levels; it improves robustness metrics relative to single-shot BRL and BRL+$α$ baselines. An ablation over LLM backbones (GPT-5, GPT-5 mini, DeepSeek-V3, Qwen-2.5-72B, Llama-4 Maverick) shows the pipeline is robust across models, while stronger models yield the highest effectiveness. These results indicate that LLM agents can integrate certificate-bearing control synthesis from high-level intent, enabling rapid end-to-end prototyping without sacrificing formal guarantees.

eess.SY↗