arXiv subjects

Jun Chen

Publications and source records attributed to Jun Chen.

At least 145 records · Page 8Linked to original sources

arXiv2025

AU-IQA: A Benchmark Dataset for Perceptual Quality Assessment of AI-Enhanced User-Generated Content

AI-based image enhancement techniques have been widely adopted in various visual applications, significantly improving the perceptual quality of user-generated content (UGC). However, the lack of specialized quality assessment models has become a significant limiting factor in this field, limiting user experience and hindering the advancement of enhancement methods. While perceptual quality assessment methods have shown strong performance on UGC and AIGC individually, their effectiveness on AI-enhanced UGC (AI-UGC) which blends features from both, remains largely unexplored. To address this gap, we construct AU-IQA, a benchmark dataset comprising 4,800 AI-UGC images produced by three representative enhancement types which include super-resolution, low-light enhancement, and denoising. On this dataset, we further evaluate a range of existing quality assessment models, including traditional IQA methods and large multimodal models. Finally, we provide a comprehensive analysis of how well current approaches perform in assessing the perceptual quality of AI-UGC. The access link to the AU-IQA is https://github.com/WNNGGU/AU-IQA-Dataset.

cs.CV↗

arXiv2025

CLIMD: A Curriculum Learning Framework for Imbalanced Multimodal Diagnosis

Clinicians usually combine information from multiple sources to achieve the most accurate diagnosis, and this has sparked increasing interest in leveraging multimodal deep learning for diagnosis. However, in real clinical scenarios, due to differences in incidence rates, multimodal medical data commonly face the issue of class imbalance, which makes it difficult to adequately learn the features of minority classes. Most existing methods tackle this issue with resampling or loss reweighting, but they are prone to overfitting or underfitting and fail to capture cross-modal interactions. Therefore, we propose a Curriculum Learning framework for Imbalanced Multimodal Diagnosis (CLIMD). Specifically, we first design multimodal curriculum measurer that combines two indicators, intra-modal confidence and inter-modal complementarity, to enable the model to focus on key samples and gradually adapt to complex category distributions. Additionally, a class distribution-guided training scheduler is introduced, which enables the model to progressively adapt to the imbalanced class distribution during training. Extensive experiments on multiple multimodal medical datasets demonstrate that the proposed method outperforms state-of-the-art approaches across various metrics and excels in handling imbalanced multimodal medical data. Furthermore, as a plug-and-play CL framework, CLIMD can be easily integrated into other models, offering a promising path for improving multimodal disease diagnosis accuracy. Code is publicly available at https://github.com/KHan-UJS/CLIMD.

cs.CV↗

arXiv2025

Algebraic Morse theory via Homological Perturbation Lemma

As a generalization of the classical killing-contractible-complexes lemma, we present algebraic Morse theory via homological perturbation lemma, in a form more general than existing presentations in the literature. Two-sided Anick resolutions due to E.~Sköldberg are generalised to algebras given by quivers with relations and a minimality criterion is provided as well. Two applications of algebraic Morse theory are presented. It is shown that the Chinese algebra of rank $n\geq 1$ is homologically smooth and of global dimension $\frac{n(n+1)}{2}$, and the minimal two-sided projective resolution of a Koszul algebra is constructed.

math.KT↗

arXiv2025

How does Labeling Error Impact Contrastive Learning? A Perspective from Data Dimensionality Reduction

In recent years, contrastive learning has achieved state-of-the-art performance in the territory of self-supervised representation learning. Many previous works have attempted to provide the theoretical understanding underlying the success of contrastive learning. Almost all of them rely on a default assumption, i.e., the label consistency assumption, which may not hold in practice (the probability of failure is called labeling error) due to the strength and randomness of common augmentation strategies, such as random resized crop (RRC). This paper investigates the theoretical impact of labeling error on the downstream classification performance of contrastive learning. We first reveal several significant negative impacts of labeling error on downstream classification risk. To mitigate these impacts, data dimensionality reduction method (e.g., singular value decomposition, SVD) is applied on original data to reduce false positive samples, and establish both theoretical and empirical evaluations. Moreover, it is also found that SVD acts as a double-edged sword, which may lead to the deterioration of downstream classification accuracy due to the reduced connectivity of the augmentation graph. Based on the above observations, we give the augmentation suggestion that we should use some moderate embedding dimension (such as $512, 1024$ in our experiments), data inflation, weak augmentation, and SVD to ensure large graph connectivity and small labeling error to improve model performance.

stat.ML↗

arXiv2025

Clustering-Guided Multi-Layer Contrastive Representation Learning for Citrus Disease Classification

Citrus, as one of the most economically important fruit crops globally, suffers severe yield depressions due to various diseases. Accurate disease detection and classification serve as critical prerequisites for implementing targeted control measures. Recent advancements in artificial intelligence, particularly deep learning-based computer vision algorithms, have substantially decreased time and labor requirements while maintaining the accuracy of detection and classification. Nevertheless, these methods predominantly rely on massive, high-quality annotated training examples to attain promising performance. By introducing two key designs: contrasting with cluster centroids and a multi-layer contrastive training (MCT) paradigm, this paper proposes a novel clustering-guided self-supervised multi-layer contrastive representation learning (CMCRL) algorithm. The proposed method demonstrates several advantages over existing counterparts: (1) optimizing with massive unannotated samples; (2) effective adaptation to the symptom similarity across distinct citrus diseases; (3) hierarchical feature representation learning. The proposed method achieves state-of-the-art performance on the public citrus image set CDD, outperforming existing methods by 4.5\%-30.1\% accuracy. Remarkably, our method narrows the performance gap with fully supervised counterparts (all samples are labeled). Beyond classification accuracy, our method shows great performance on other evaluation metrics (F1 score, precision, and recall), highlighting the robustness against the class imbalance challenge.

cs.CV↗

arXiv2025

Diffusion-Based Imaginative Coordination for Bimanual Manipulation

Bimanual manipulation is crucial in robotics, enabling complex tasks in industrial automation and household services. However, it poses significant challenges due to the high-dimensional action space and intricate coordination requirements. While video prediction has been recently studied for representation learning and control, leveraging its ability to capture rich dynamic and behavioral information, its potential for enhancing bimanual coordination remains underexplored. To bridge this gap, we propose a unified diffusion-based framework for the joint optimization of video and action prediction. Specifically, we propose a multi-frame latent prediction strategy that encodes future states in a compressed latent space, preserving task-relevant features. Furthermore, we introduce a unidirectional attention mechanism where video prediction is conditioned on the action, while action prediction remains independent of video prediction. This design allows us to omit video prediction during inference, significantly enhancing efficiency. Experiments on two simulated benchmarks and a real-world setting demonstrate a significant improvement in the success rate over the strong baseline ACT using our method, achieving a \textbf{24.9\%} increase on ALOHA, an \textbf{11.1\%} increase on RoboTwin, and a \textbf{32.5\%} increase in real-world experiments. Our models and code are publicly available at https://github.com/return-sleep/Diffusion_based_imaginative_Coordination.

cs.RO↗

arXiv2025

Benchmarking Generalizable Bimanual Manipulation: RoboTwin Dual-Arm Collaboration Challenge at CVPR 2025 MEIS Workshop

Embodied Artificial Intelligence (Embodied AI) is an emerging frontier in robotics, driven by the need for autonomous systems that can perceive, reason, and act in complex physical environments. While single-arm systems have shown strong task performance, collaborative dual-arm systems are essential for handling more intricate tasks involving rigid, deformable, and tactile-sensitive objects. To advance this goal, we launched the RoboTwin Dual-Arm Collaboration Challenge at the 2nd MEIS Workshop, CVPR 2025. Built on the RoboTwin Simulation platform (1.0 and 2.0) and the AgileX COBOT-Magic Robot platform, the competition consisted of three stages: Simulation Round 1, Simulation Round 2, and a final Real-World Round. Participants totally tackled 17 dual-arm manipulation tasks, covering rigid, deformable, and tactile-based scenarios. The challenge attracted 64 global teams and over 400 participants, producing top-performing solutions like SEM and AnchorDP3 and generating valuable insights into generalizable bimanual policy learning. This report outlines the competition setup, task design, evaluation methodology, key findings and future direction, aiming to support future research on robust and generalizable bimanual manipulation policies. The Challenge Webpage is available at https://robotwin-benchmark.github.io/cvpr-2025-challenge/.

cs.RO↗

arXiv2025

Source-Channel Separation Theorems for Distortion Perception Coding

It is well known that separation between lossy source coding and channel coding is asymptotically optimal under classical additive distortion measures. Recently, coding under a new class of quality considerations, often referred to as perception or realism, has attracted significant attention due to its close connection to neural generative models and semantic communications. In this work, we revisit source-channel separation under the consideration of distortion-perception. We show that when the perception quality is measured on the block level, i.e., in the strong-sense, the optimality of separation still holds when common randomness is shared between the encoder and the decoder; however, separation is no longer optimal when such common randomness is not available. In contrast, when the perception quality is the average per-symbol measure, i.e., in the weak-sense, the optimality of separation holds regardless of the availability of common randomness.

cs.IT↗

arXiv2025

Establishment of global phase coherence in a highly disordered fractal MgO/MgB2 nanocomposite: Roles of interface, morphology and defect

Recently, we have reported that a highly disordered fractal MgO/MgB2 nanocomposite exhibits bulk-like superconducting properties with isotropic pinning, showing an excellent phase-coherent capability irrespective of the low volume fraction (~30 vol. %) of MgB2 [Uchino et al., Phys. Rev. B 101, 035146 (2020); Teramachi et al,, Phys. Rev. B 108, 155146 (2023)]. Hence, this nanocomposite provides a useful experimental system to investigate the relationship between the structural disorder and the establishment of the superconducting phase coherence. In this work, we show from 3D focused ion beam scanning electron microscopy (FIB-SEM) data that in the nanocomposite, a complex MgO/MgB2 microstructure spreads isotropically throughout the sample with a constant fractal dimension of ~1.67. Atomic-resolution scanning transmission electron microscopy (STEM) has revealed that the MgO/MgB2 interfaces are atomically clean and free from amorphous grain boundaries, even leading to atomically coherent interfaces. Detailed ac susceptibility measurements have demonstrated a smooth crossover from an intragranular to an intergranular superconducting regime, giving evidence of the establishment of the critical state due to strong intergranular coupling just below the superconducting transition temperature. Also, spatially-resolved cathodoluminescence measurements have demonstrated that oxygen vacancies in the MgO-rich phase tend to aggregate near the MgO/MgB2 boundary regions, forming long channels of oxygen vacancies through the nanocomposite. These channels of oxygen vacancies will contribute to the long-range carrier transfer and the related Andreev reflection via coherent tunneling of charge carriers among the oxygen vacancy sites.

cond-mat.supr-con↗

arXiv2025

MoiréXNet: Adaptive Multi-Scale Demoiréing with Linear Attention Test-Time Training and Truncated Flow Matching Prior

This paper introduces a novel framework for image and video demoiréing by integrating Maximum A Posteriori (MAP) estimation with advanced deep learning techniques. Demoiréing addresses inherently nonlinear degradation processes, which pose significant challenges for existing methods. Traditional supervised learning approaches either fail to remove moiré patterns completely or produce overly smooth results. This stems from constrained model capacity and scarce training data, which inadequately represent the clean image distribution and hinder accurate reconstruction of ground-truth images. While generative models excel in image restoration for linear degradations, they struggle with nonlinear cases such as demoiréing and often introduce artifacts. To address these limitations, we propose a hybrid MAP-based framework that integrates two complementary components. The first is a supervised learning model enhanced with efficient linear attention Test-Time Training (TTT) modules, which directly learn nonlinear mappings for RAW-to-sRGB demoiréing. The second is a Truncated Flow Matching Prior (TFMP) that further refines the outputs by aligning them with the clean image distribution, effectively restoring high-frequency details and suppressing artifacts. These two components combine the computational efficiency of linear attention with the refinement abilities of generative models, resulting in improved restoration performance.

cs.CV↗

arXiv2025

NTIRE 2025 Image Shadow Removal Challenge Report

This work examines the findings of the NTIRE 2025 Shadow Removal Challenge. A total of 306 participants have registered, with 17 teams successfully submitting their solutions during the final evaluation phase. Following the last two editions, this challenge had two evaluation tracks: one focusing on reconstruction fidelity and the other on visual perception through a user study. Both tracks were evaluated with images from the WSRD+ dataset, simulating interactions between self- and cast-shadows with a large number of diverse objects, textures, and materials.

cs.CV↗

arXiv2025

PNCS:Power-Norm Cosine Similarity for Diverse Client Selection in Federated Learning

Federated Learning (FL) has emerged as a powerful paradigm for leveraging diverse datasets from multiple sources while preserving data privacy by avoiding centralized storage. However, many existing approaches fail to account for the intricate gradient correlations between remote clients, a limitation that becomes especially problematic in data heterogeneity scenarios. In this work, we propose a novel FL framework utilizing Power-Norm Cosine Similarity (PNCS) to improve client selection for model aggregation. By capturing higher-order gradient moments, PNCS addresses non-IID data challenges, enhancing convergence speed and accuracy. Additionally, we introduce a simple algorithm ensuring diverse client selection through a selection history queue. Experiments with a VGG16 model across varied data partitions demonstrate consistent improvements over state-of-the-art methods.

cs.LG↗

arXiv2025

LLaSE-G1: Incentivizing Generalization Capability for LLaMA-based Speech Enhancement

Recent advancements in language models (LMs) have demonstrated strong capabilities in semantic understanding and contextual modeling, which have flourished in generative speech enhancement (SE). However, many LM-based SE approaches primarily focus on semantic information, often neglecting the critical role of acoustic information, which leads to acoustic inconsistency after enhancement and limited generalization across diverse SE tasks. In this paper, we introduce LLaSE-G1, a LLaMA-based language model that incentivizes generalization capabilities for speech enhancement. LLaSE-G1 offers the following key contributions: First, to mitigate acoustic inconsistency, LLaSE-G1 employs continuous representations from WavLM as input and predicts speech tokens from X-Codec2, maximizing acoustic preservation. Second, to promote generalization capability, LLaSE-G1 introduces dual-channel inputs and outputs, unifying multiple SE tasks without requiring task-specific IDs. Third, LLaSE-G1 outperforms prior task-specific discriminative and generative SE models, demonstrating scaling effects at test time and emerging capabilities for unseen SE tasks. Additionally, we release our code and models to support further research in this area.

eess.AS↗

arXiv2025

Decentralized Optimization on Compact Submanifolds by Quantized Riemannian Gradient Tracking

This paper considers the problem of decentralized optimization on compact submanifolds, where a finite sum of smooth (possibly non-convex) local functions is minimized by $n$ agents forming an undirected and connected graph. However, the efficiency of distributed optimization is often hindered by communication bottlenecks. To mitigate this, we propose the Quantized Riemannian Gradient Tracking (Q-RGT) algorithm, where agents update their local variables using quantized gradients. The introduction of quantization noise allows our algorithm to bypass the constraints of the accurate Riemannian projection operator (such as retraction), further improving iterative efficiency. To the best of our knowledge, this is the first algorithm to achieve an $\mathcal{O}(1/K)$ convergence rate in the presence of quantization, matching the convergence rate of methods without quantization. Additionally, we explicitly derive lower bounds on decentralized consensus associated with a function of quantization levels. Numerical experiments demonstrate that Q-RGT performs comparably to non-quantized methods while reducing communication bottlenecks and computational overhead.

math.OC↗

arXiv2025

FlowSE: Efficient and High-Quality Speech Enhancement via Flow Matching

Generative models have excelled in audio tasks using approaches such as language models, diffusion, and flow matching. However, existing generative approaches for speech enhancement (SE) face notable challenges: language model-based methods suffer from quantization loss, leading to compromised speaker similarity and intelligibility, while diffusion models require complex training and high inference latency. To address these challenges, we propose FlowSE, a flow-matching-based model for SE. Flow matching learns a continuous transformation between noisy and clean speech distributions in a single pass, significantly reducing inference latency while maintaining high-quality reconstruction. Specifically, FlowSE trains on noisy mel spectrograms and optional character sequences, optimizing a conditional flow matching loss with ground-truth mel spectrograms as supervision. It implicitly learns speech's temporal-spectral structure and text-speech alignment. During inference, FlowSE can operate with or without textual information, achieving impressive results in both scenarios, with further improvements when transcripts are available. Extensive experiments demonstrate that FlowSE significantly outperforms state-of-the-art generative methods, establishing a new paradigm for generative-based SE and demonstrating the potential of flow matching to advance the field. Our code, pre-trained checkpoints, and audio samples are available.

eess.AS↗

arXiv2025

Reinforcement Learning from User Feedback

As large language models (LLMs) are increasingly deployed in diverse user facing applications, aligning them with real user preferences becomes essential. Existing methods like Reinforcement Learning from Human Feedback (RLHF) rely on expert annotators trained on manually defined guidelines, whose judgments may not reflect the priorities of everyday users. We introduce Reinforcement Learning from User Feedback (RLUF), a framework for aligning LLMs directly to implicit signals from users in production. RLUF addresses key challenges of user feedback: user feedback is often binary (e.g., emoji reactions), sparse, and occasionally adversarial. We train a reward model, P[Love], to predict the likelihood that an LLM response will receive a Love Reaction, a lightweight form of positive user feedback, and integrate P[Love] into a multi-objective policy optimization framework alongside helpfulness and safety objectives. In large-scale experiments, we show that P[Love] is predictive of increased positive feedback and serves as a reliable offline evaluator of future user behavior. Policy optimization using P[Love] significantly raises observed positive-feedback rates, including a 28% increase in Love Reactions during live A/B tests. However, optimizing for positive reactions introduces reward hacking challenges, requiring careful balancing of objectives. By directly leveraging implicit signals from users, RLUF offers a path to aligning LLMs with real-world user preferences at scale.

cs.AI↗

arXiv2025

Crescent domain-wall pairs in anisotropic two-dimensional MnOI monolayer with fast dynamics

The controllable and efficient manipulation is always a key challenge for the application of topological magnetic textures in spintronic devices. By first-principles calculations and atomistic simulations, the present work reveals an exotic domain-wall pair in crescent shape, which possesses topology in one dimension and particle-like robustness in two dimensions. Its antiferromagnetic version is observed in the two-dimensional MnOI monolayer with a strong anisotropy, and the ferromagnetic version exists under an appropriate strain. Both of them can be driven efficiently by current to move in a straight line along a certain direction. Hereby, another possible path is provided to realize the application of topological magnetic texture in the next-generation high-speed spintronic devices.

cond-mat.mtrl-sci↗

arXiv2025

Tracking an eruptive intermediate prominence originating from the farside of the Sun

In this paper, we carry out multiwavelength and multiview observations of the eruption of an intermediate prominence originating from the farside of the Sun on 2023 March 12. The southeast footpoint of the prominence is located in active region (AR) 13252. The eruption generates a B7.8 class flare and a partial halo coronal mass ejection (CME). The prominence takes off at 02:00 UT and accelerates for nearly three hours. Rotation of the southeast leg of the prominence in the counterclockwise direction is revealed by spectroscopic and imaging observations. The apex of the prominence changes from a smooth loop to a cusp structure during the rising motion and the northwest leg displays a drift motion after 04:30 UT, implying a writhing motion. Hence, the prominence eruption is most likely triggered by ideal kink instability. For the first time, we apply the Graduated Cylindrical Shell (GCS) modeling in three-dimensional reconstruction and tracking of the prominence for nearly two hours. Both the source region (110$\degr$E, 43$\degr$N) and northwest footpoint (162$\degr$E, 44$\degr$N) are located. The edge-on and face-on angular widths of the prominence are $\sim$6$\degr$ and $\sim$86$\degr$, respectively. The axis has a tilt angle of $\sim$70$\degr$ with the meridian. The heliocentric distance of the prominence leading edge increases from $\sim$1.26\,$R_{\sun}$ to $\sim$2.27\,$R_{\sun}$. The true speed of the CME increases from $\sim$610 to $\sim$849 km s$^{-1}$.

astro-ph.SR↗