arXiv subjects

Chen Xu

Publications and source records attributed to Chen Xu.

At least 91 records · Page 5Linked to original sources

arXiv2025

Triad: Empowering LMM-based Anomaly Detection with Vision Expert-guided Visual Tokenizer and Manufacturing Process

Although recent methods have tried to introduce large multimodal models (LMMs) into industrial anomaly detection (IAD), their generalization in the IAD field is far inferior to that for general purposes. We summarize the main reasons for this gap into two aspects. On one hand, general-purpose LMMs lack cognition of defects in the visual modality, thereby failing to sufficiently focus on defect areas. Therefore, we propose to modify the AnyRes structure of the LLaVA model, providing the potential anomalous areas identified by existing IAD models to the LMMs. On the other hand, existing methods mainly focus on identifying defects by learning defect patterns or comparing with normal samples, yet they fall short of understanding the causes of these defects. Considering that the generation of defects is closely related to the manufacturing process, we propose a manufacturing-driven IAD paradigm. An instruction-tuning dataset for IAD (InstructIAD) and a data organization approach for Chain-of-Thought with manufacturing (CoT-M) are designed to leverage the manufacturing process for IAD. Based on the above two modifications, we present Triad, a novel LMM-based method incorporating an expert-guided region-of-interest tokenizer and manufacturing process for industrial anomaly detection. Extensive experiments show that our Triad not only demonstrates competitive performance against current LMMs but also achieves further improved accuracy when equipped with manufacturing processes. Source code, training data, and pre-trained models will be publicly available at https://github.com/tzjtatata/Triad.

cs.CV↗

arXiv2025

MQuant: Unleashing the Inference Potential of Multimodal Large Language Models via Full Static Quantization

Multimodal large language models (MLLMs) have garnered widespread attention due to their ability to understand multimodal input. However, their large parameter sizes and substantial computational demands severely hinder their practical deployment and application.While quantization is an effective way to reduce model size and inference latency, its application to MLLMs remains underexplored. In this paper, we propose MQuant, a post-training quantization (PTQ) framework designed to tackle the unique challenges of multimodal large language models (MLLMs). Conventional quantization often struggles with MLLMs because of (a) high inference latency from large visual token counts, (b) distributional disparities between visual and textual tokens, and (c) extreme outliers introduced by Hadamard-based transformations. To address these issues, MQuant introduces: Modality-Specific Static Quantization (MSQ), assigning distinct static scales for visual vs. textual tokens; Attention-Invariant Flexible Switching (AIFS), reordering tokens to preserve casual attention while eliminating expensive token-wise scale computations; Rotation Magnitude Suppression (RMS), mitigating weight outliers arising from online Hadamard rotations. On five mainstream MLLMs (including Qwen-VL, MiniCPM-V, CogVLM2), MQuant under W4A8 achieves near-floating-point accuracy (<1% degradation) while reducing inference latency by up to 30%, significantly outperforming existing PTQ baselines. Our MQuant effectively bridges the gap for efficient and accurate MLLMs inference in resource-constrained devices. Code has been released in https://github.com/StiphyJay/MQuant.

cs.CV↗

arXiv2025

Bridging Simulation and Experiment: A Self-Supervised Domain Adaptation Framework for Concrete Damage Classification

Reliable assessment of concrete degradation is critical for ensuring structural safety and longevity of engineering structures. This study proposes a self-supervised domain adaptation framework for robust concrete damage classification using coda wave signals. To support this framework, an advanced virtual testing platform is developed, combining multiscale modeling of concrete degradation with ultrasonic wave propagation simulations. This setup enables the generation of large-scale labeled synthetic data under controlled conditions, reducing the dependency on costly and time-consuming experimental labeling. However, neural networks trained solely on synthetic data often suffer from degraded performance when applied to experimental data due to domain shifts. To bridge this domain gap, the proposed framework integrates domain adversarial training, minimum class confusion loss, and the Bootstrap Your Own Latent (BYOL) strategy. These components work jointly to facilitate effective knowledge transfer from the labeled simulation domain to the unlabeled experimental domain, achieving accurate and reliable damage classification in concrete. Extensive experiments demonstrate that the proposed method achieves notable performance improvements, reaching an accuracy of 0.7762 and a macro F1 score of 0.7713, outperforming both the plain 1D CNN baseline and six representative domain adaptation techniques. Moreover, the method exhibits high robustness across training runs and introduces only minimal additional computational cost. These findings highlight the practical potential of the proposed simulation-driven and label-efficient framework for real-world applications in structural health monitoring.

cs.CE↗

arXiv2025

Sliding two-dimensional superconductivity and charge-density-wave state in a bulk crystal

Superconductivity in the two-dimensional (2D) limit is a fertile ground for exotic quantum phenomena-many of which remain elusive in their 3D counterparts. While studies of 2D superconductivity have predominantly focused on mono- or few-layer systems, we demonstrate an alternative route-interlayer sliding in bulk crystals. Through a precisely controlled growth strategy, we engineer interlayer sliding in bulk 3R-NbSe2, deliberately disrupting [001] mirror symmetry and drastically suppressing interlayer coupling. Remarkably, this structural manipulation stabilizes Ising-type superconductivity coexisting with an unconventional charge-density-wave (CDW) state akin to that of monolayer 2H-NbSe2. The sliding phase exhibits a pronounced suppression of the upper critical field at low temperatures, revealing a delicate competition between Ising and Rashba spin-orbit coupling (SOC) in the globally noncentrosymmetric lattice. Intriguingly, the superconducting state displays two-fold symmetry, a signature that may arise from asymmetric SOC or a multi-component pairing order parameter. Our work establishes interlayer sliding as a symmetry-breaking tool to promote 2D superconductivity in bulk materials-without resorting to extrinsic intercalation or doping. More broadly, this approach sets a paradigm for unlocking hidden quantum states in layered materials, offering a new dimension in design of quantum matter.

cond-mat.supr-con↗

arXiv2025

Step-3 is Large yet Affordable: Model-system Co-design for Cost-effective Decoding

Large language models (LLMs) face low hardware efficiency during decoding, especially for long-context reasoning tasks. This paper introduces Step-3, a 321B-parameter VLM with hardware-aware model-system co-design optimized for minimizing decoding costs. Step-3 innovates in two key dimensions: (1) A novel Multi-Matrix Factorization Attention (MFA) mechanism that significantly reduces both KV cache size and computation while maintaining high attention expressiveness, and (2) Attention-FFN Disaggregation (AFD), a distributed inference system that decouples attention and Feed-Forward Network (FFN) layers into specialized subsystems. This co-design achieves unprecedented cost efficiency: Step-3 significantly reduces theoretical decoding costs compared with models like DeepSeek-V3 and Qwen3 MoE 235B, with the gains widening at longer context. Step-3 achieves low cost while activating 38B parameters per token (more than DeepSeek-V3 and Qwen3 MoE 235B), demonstrating that hardware-aligned attention arithmetic intensity, MoE sparsity, and AFD are critical to cost-effectiveness. We perform a head-to-head comparison with DeepSeek-V3 in its favorable scenarios. Our implementation on Hopper GPUs achieves a decoding throughput of up to 4,039 tokens per second per GPU under 50ms TPOT SLA (4K context, FP8, no MTP). It is higher than DeepSeek-V3's 2,324 in the same setup and sets a new Pareto frontier for LLM decoding.

cs.LG↗

arXiv2025

Prototypical Progressive Alignment and Reweighting for Generalizable Semantic Segmentation

Generalizable semantic segmentation aims to perform well on unseen target domains, a critical challenge due to real-world applications requiring high generalizability. Class-wise prototypes, representing class centroids, serve as domain-invariant cues that benefit generalization due to their stability and semantic consistency. However, this approach faces three challenges. First, existing methods often adopt coarse prototypical alignment strategies, which may hinder performance. Second, naive prototypes computed by averaging source batch features are prone to overfitting and may be negatively affected by unrelated source data. Third, most methods treat all source samples equally, ignoring the fact that different features have varying adaptation difficulties. To address these limitations, we propose a novel framework for generalizable semantic segmentation: Prototypical Progressive Alignment and Reweighting (PPAR), leveraging the strong generalization ability of the CLIP model. Specifically, we define two prototypes: the Original Text Prototype (OTP) and Visual Text Prototype (VTP), generated via CLIP to serve as a solid base for alignment. We then introduce a progressive alignment strategy that aligns features in an easy-to-difficult manner, reducing domain gaps gradually. Furthermore, we propose a prototypical reweighting mechanism that estimates the reliability of source data and adjusts its contribution, mitigating the effect of irrelevant or harmful features (i.e., reducing negative transfer). We also provide a theoretical analysis showing the alignment between our method and domain generalization theory. Extensive experiments across multiple benchmarks demonstrate that PPAR achieves state-of-the-art performance, validating its effectiveness.

cs.CV↗

arXiv2025

Wehrl Entropy and Entanglement Complexity of Quantum Spin Systems

The Wehrl entropy of a quantum state is the Shannon entropy of its coherent-state distribution function, and remains non-zero even for pure states. We investigate the relationship between this entropy and the many-particle quantum entanglement, for $N$ spin-1/2 particles. Explicitly, we numerically calculate the Wehrl entropy of various $N$-particle ($2\leq N\leq 20$) entangled pure states, with respect to the SU(2)$^{\otimes N}$ coherent states. Our results show that for the large-$N$ ($N\gtrsim 10$) systems the Wehrl entropy of the highly chaotic entangled states (e.g., $2^{-N/2}\sum_{s_1,s_2,...,s_N=\uparrow,\downarrow}|s_1,s_2,...,s_N\rangle e^{-iϕ_{s_1,s_2,...,s_N}}$, with $ϕ_{s_1,s_2,...,s_N}$ being random angles) are substantially larger than that of the very regular entangled states (e.g., the Greenberger-Horne-Zeilinger state). Therefore, the Wehrl entropy can reflect the complexity of the quantum entanglement of many-body pure states, as proposed by A. Sugita (Jour. Phys. A 36, 9081 (2003)). In particular, the Wehrl entropy per particle (WEPP) can be used as a quantitative description of this entanglement complexity. Unlike other quantities used to evaluate this complexity (e.g., the degree of entanglement between a subsystem and the other particles), the WEPP does not necessitate the division of the total system into two subsystems. We further demonstrate that many-body pure entangled states can be classified into three types, based on the behavior of the WEPP in the limit $N \rightarrow \infty$: states approaching that of a maximally mixed state, those approaching completely separable pure states, and a third category lying between these two extremes. Each type exhibits fundamentally different entanglement complexity.

cond-mat.stat-mech↗

arXiv2025

Bayesian Interpretation of Husimi Function and Wehrl Entropy

Husimi function (Q-function) of a quantum state is the distribution function of the density operator in the coherent state representation. It is widely used in theoretical research, such as in quantum optics. The Wehrl entropy is the Shannon entropy of the Husimi function, and is non-zero even for pure states. This entropy has been extensively studied in mathematical physics. Recent research also suggests a significant connection between the Wehrl entropy and many-body quantum entanglement in spin systems. We investigate the statistical interpretation of the Husimi function and the Wehrl entropy, taking the system of $N$ spin-1/2 particles as an example. Due to the completeness of coherent states, the Husimi function and Wehrl entropy can be explained via the positive operator-valued measurement (POVM) theory, although the coherent states are not a set of orthonormal basis. Here, with the help of the Bayes' theorem, we provide an alternative probabilistic interpretation for the Husimi function and the Wehrl entropy. This interpretation is based on direct measurements of the system, and thus does not require the introduction of an ancillary system as in POVM theory. Moreover, under this interpretation the classical correspondences of the Husimi function and Wehrl entropy are just phase-space probability distribution function of $N$ classical tops, and its associated entropy, respectively. Therefore, this explanation contributes to a better understanding of the relationship between the Husimi function, Wehrl entropy, and classical-quantum correspondence. The generalization of this statistical interpretation to continuous-variable systems is also discussed.

quant-ph↗

arXiv2025

A Careful Examination of Large Behavior Models for Multitask Dexterous Manipulation

Robot manipulation has seen tremendous progress in recent years, with imitation learning policies enabling successful performance of dexterous and hard-to-model tasks. Concurrently, scaling data and model size has led to the development of capable language and vision foundation models, motivating large-scale efforts to create general-purpose robot foundation models. While these models have garnered significant enthusiasm and investment, meaningful evaluation of real-world performance remains a challenge, limiting both the pace of development and inhibiting a nuanced understanding of current capabilities. In this paper, we rigorously evaluate multitask robot manipulation policies, referred to as Large Behavior Models (LBMs), by extending the Diffusion Policy paradigm across a corpus of simulated and real-world robot data. We propose and validate an evaluation pipeline to rigorously analyze the capabilities of these models with statistical confidence. We compare against single-task baselines through blind, randomized trials in a controlled setting, using both simulation and real-world experiments. We find that multi-task pretraining makes the policies more successful and robust, and enables teaching complex new tasks more quickly, using a fraction of the data when compared to single-task baselines. Moreover, performance predictably increases as pretraining scale and diversity grows. Project page: https://toyotaresearchinstitute.github.io/lbm1/

cs.RO↗

arXiv2025

Can We Detect Failures Without Failure Data? Uncertainty-Aware Runtime Failure Detection for Imitation Learning Policies

Recent years have witnessed impressive robotic manipulation systems driven by advances in imitation learning and generative modeling, such as diffusion- and flow-based approaches. As robot policy performance increases, so does the complexity and time horizon of achievable tasks, inducing unexpected and diverse failure modes that are difficult to predict a priori. To enable trustworthy policy deployment in safety-critical human environments, reliable runtime failure detection becomes important during policy inference. However, most existing failure detection approaches rely on prior knowledge of failure modes and require failure data during training, which imposes a significant challenge in practicality and scalability. In response to these limitations, we present FAIL-Detect, a modular two-stage approach for failure detection in imitation learning-based robotic manipulation. To accurately identify failures from successful training data alone, we frame the problem as sequential out-of-distribution (OOD) detection. We first distill policy inputs and outputs into scalar signals that correlate with policy failures and capture epistemic uncertainty. FAIL-Detect then employs conformal prediction (CP) as a versatile framework for uncertainty quantification with statistical guarantees. Empirically, we thoroughly investigate both learned and post-hoc scalar signal candidates on diverse robotic manipulation tasks. Our experiments show learned signals to be mostly consistently effective, particularly when using our novel flow-based density estimator. Furthermore, our method detects failures more accurately and faster than state-of-the-art (SOTA) failure detection baselines. These results highlight the potential of FAIL-Detect to enhance the safety and reliability of imitation learning-based robotic systems as they progress toward real-world deployment.

cs.RO↗

arXiv2025

Exploiting Both Pilots and Data Payloads for Integrated Sensing and Communications

Integrated sensing and communications (ISAC) is one of the key enabling technologies in future sixth-generation (6G) networks. Current ISAC systems predominantly rely on deterministic pilot signals within the signal frame to accomplish sensing tasks. However, these pilot signals typically occupy only a small portion, e.g., 0.15% to 25%, of the time-frequency resources. To enhance the system utility, a promising solution is to repurpose the extensive random data payload signals for sensing tasks. In this paper, we analyze the ISAC performance of a multi-antenna system where both deterministic pilot and random data symbols are employed for sensing tasks. By capitalizing on random matrix theory (RMT), we first derive a semi-closed-form asymptotic expression of the ergodic linear minimum mean square error (ELMMSE). Then, we formulate an ISAC precoding optimization problem to minimize the ELMMSE, which is solved via a specifically tailored successive convex approximation (SAC) algorithm. To provide system insights, we further derive a closed-form expression for the asymptotic ELMMSE at high signal-to-noise ratios (SNRs). Our analysis reveals that, compared with conventional sensing implemented by deterministic signals, the sensing performance degradation induced by random signals is critically determined by the ratio of the transmit antenna size to the data symbol length. Based on this result, the ISAC precoding optimization problem at high SNRs is transformed into a convex optimization problem that can be efficiently solved. Simulation results validate the accuracy of the derived asymptotic expressions of ELMMSE and the performance of the proposed precoding schemes. Particularly, by leveraging data payload signals for sensing tasks, the sensing error is reduced by up to 5.6 dB compared to conventional pilot-based sensing.

eess.SP↗

arXiv2025

Step-Audio-AQAA: a Fully End-to-End Expressive Large Audio Language Model

Large Audio-Language Models (LALMs) have significantly advanced intelligent human-computer interaction, yet their reliance on text-based outputs limits their ability to generate natural speech responses directly, hindering seamless audio interactions. To address this, we introduce Step-Audio-AQAA, a fully end-to-end LALM designed for Audio Query-Audio Answer (AQAA) tasks. The model integrates a dual-codebook audio tokenizer for linguistic and semantic feature extraction, a 130-billion-parameter backbone LLM and a neural vocoder for high-fidelity speech synthesis. Our post-training approach employs interleaved token-output of text and audio to enhance semantic coherence and combines Direct Preference Optimization (DPO) with model merge to improve performance. Evaluations on the StepEval-Audio-360 benchmark demonstrate that Step-Audio-AQAA excels especially in speech control, outperforming the state-of-art LALMs in key areas. This work contributes a promising solution for end-to-end LALMs and highlights the critical role of token-based vocoder in enhancing overall performance for AQAA tasks.

cs.SD↗

arXiv2025

Precise Zero-Shot Pointwise Ranking with LLMs through Post-Aggregated Global Context Information

Recent advancements have successfully harnessed the power of Large Language Models (LLMs) for zero-shot document ranking, exploring a variety of prompting strategies. Comparative approaches like pairwise and listwise achieve high effectiveness but are computationally intensive and thus less practical for larger-scale applications. Scoring-based pointwise approaches exhibit superior efficiency by independently and simultaneously generating the relevance scores for each candidate document. However, this independence ignores critical comparative insights between documents, resulting in inconsistent scoring and suboptimal performance. In this paper, we aim to improve the effectiveness of pointwise methods while preserving their efficiency through two key innovations: (1) We propose a novel Global-Consistent Comparative Pointwise Ranking (GCCP) strategy that incorporates global reference comparisons between each candidate and an anchor document to generate contrastive relevance scores. We strategically design the anchor document as a query-focused summary of pseudo-relevant candidates, which serves as an effective reference point by capturing the global context for document comparison. (2) These contrastive relevance scores can be efficiently Post-Aggregated with existing pointwise methods, seamlessly integrating essential Global Context information in a training-free manner (PAGC). Extensive experiments on the TREC DL and BEIR benchmark demonstrate that our approach significantly outperforms previous pointwise methods while maintaining comparable efficiency. Our method also achieves competitive performance against comparative methods that require substantially more computational resources. More analyses further validate the efficacy of our anchor construction strategy.

cs.IR↗

arXiv2025

Collapsing Sequence-Level Data-Policy Coverage via Poisoning Attack in Offline Reinforcement Learning

Offline reinforcement learning (RL) heavily relies on the coverage of pre-collected data over the target policy's distribution. Existing studies aim to improve data-policy coverage to mitigate distributional shifts, but overlook security risks from insufficient coverage, and the single-step analysis is not consistent with the multi-step decision-making nature of offline RL. To address this, we introduce the sequence-level concentrability coefficient to quantify coverage, and reveal its exponential amplification on the upper bound of estimation errors through theoretical analysis. Building on this, we propose the Collapsing Sequence-Level Data-Policy Coverage (CSDPC) poisoning attack. Considering the continuous nature of offline RL data, we convert state-action pairs into decision units, and extract representative decision patterns that capture multi-step behavior. We identify rare patterns likely to cause insufficient coverage, and poison them to reduce coverage and exacerbate distributional shifts. Experiments show that poisoning just 1% of the dataset can degrade agent performance by 90%. This finding provides new perspectives for analyzing and safeguarding the security of offline RL.

cs.LG↗

arXiv2025

A Survey of Automatic Evaluation Methods on Text, Visual and Speech Generations

Recent advances in deep learning have significantly enhanced generative AI capabilities across text, images, and audio. However, automatically evaluating the quality of these generated outputs presents ongoing challenges. Although numerous automatic evaluation methods exist, current research lacks a systematic framework that comprehensively organizes these methods across text, visual, and audio modalities. To address this issue, we present a comprehensive review and a unified taxonomy of automatic evaluation methods for generated content across all three modalities; We identify five fundamental paradigms that characterize existing evaluation approaches across these domains. Our analysis begins by examining evaluation methods for text generation, where techniques are most mature. We then extend this framework to image and audio generation, demonstrating its broad applicability. Finally, we discuss promising directions for future research in cross-modal evaluation methodologies.

cs.CL↗

arXiv2025

ECP-Mamba: An Efficient Multi-scale Self-supervised Contrastive Learning Method with State Space Model for PolSAR Image Classification

Recently, polarimetric synthetic aperture radar (PolSAR) image classification has been greatly promoted by deep neural networks. However,current deep learning-based PolSAR classification methods encounter difficulties due to its dependence on extensive labeled data and the computational inefficiency of architectures like Transformers. This paper presents ECP-Mamba, an efficient framework integrating multi-scale self-supervised contrastive learning with a state space model (SSM) backbone. Specifically, ECP-Mamba addresses annotation scarcity through a multi-scale predictive pretext task based on local-to-global feature correspondences, which uses a simplified self-distillation paradigm without negative sample pairs. To enhance computational efficiency,the Mamba architecture (a selective SSM) is first tailored for pixel-wise PolSAR classification task by designing a spiral scan strategy. This strategy prioritizes causally relevant features near the central pixel, leveraging the localized nature of pixel-wise classification tasks. Additionally, the lightweight Cross Mamba module is proposed to facilitates complementary multi-scale feature interaction with minimal overhead. Extensive experiments across four benchmark datasets demonstrate ECP-Mamba's effectiveness in balancing high accuracy with resource efficiency. On the Flevoland 1989 dataset, ECP-Mamba achieves state-of-the-art performance with an overall accuracy of 99.70%, average accuracy of 99.64% and Kappa coefficient of 99.62e-2. Our code will be available at https://github.com/HaixiaBi1982/ECP_Mamba.

cs.CV↗

arXiv2025

Label-shift robust federated feature screening for high-dimensional classification

Distributed and federated learning are important tools for high-dimensional classification of large datasets. To reduce computational costs and overcome the curse of dimensionality, feature screening plays a pivotal role in eliminating irrelevant features during data preprocessing. However, data heterogeneity, particularly label shifting across different clients, presents significant challenges for feature screening. This paper introduces a general framework that unifies existing screening methods and proposes a novel utility, label-shift robust federated feature screening (LR-FFS), along with its federated estimation procedure. The framework facilitates a uniform analysis of methods and systematically characterizes their behaviors under label shift conditions. Building upon this framework, LR-FFS leverages conditional distribution functions and expectations to address label shift without adding computational burdens and remains robust against model misspecification and outliers. Additionally, the federated procedure ensures computational efficiency and privacy protection while maintaining screening effectiveness comparable to centralized processing. We also provide a false discovery rate (FDR) control method for federated feature screening. Experimental results and theoretical analyses demonstrate LR-FFS's superior performance across diverse client environments, including those with varying class distributions, sample sizes, and missing categorical data.

stat.ML↗

arXiv2025

T2I-Eval-R1: Reinforcement Learning-Driven Reasoning for Interpretable Text-to-Image Evaluation

The rapid progress in diffusion-based text-to-image (T2I) generation has created an urgent need for interpretable automatic evaluation methods that can assess the quality of generated images, therefore reducing the human annotation burden. To reduce the prohibitive cost of relying on commercial models for large-scale evaluation, and to improve the reasoning capabilities of open-source models, recent research has explored supervised fine-tuning (SFT) of multimodal large language models (MLLMs) as dedicated T2I evaluators. However, SFT approaches typically rely on high-quality critique datasets, which are either generated by proprietary LLMs-with potential issues of bias and inconsistency-or annotated by humans at high cost, limiting their scalability and generalization. To address these limitations, we propose T2I-Eval-R1, a novel reinforcement learning framework that trains open-source MLLMs using only coarse-grained quality scores, thereby avoiding the need for annotating high-quality interpretable evaluation rationale. Our approach integrates Group Relative Policy Optimization (GRPO) into the instruction-tuning process, enabling models to generate both scalar scores and interpretable reasoning chains with only easy accessible annotated judgment scores or preferences. Furthermore, we introduce a continuous reward formulation that encourages score diversity and provides stable optimization signals, leading to more robust and discriminative evaluation behavior. Experimental results on three established T2I meta-evaluation benchmarks demonstrate that T2I-Eval-R1 achieves significantly higher alignment with human assessments and offers more accurate interpretable score rationales compared to strong baseline methods.

cs.AI↗