arXiv subjects

Jun Chen

Publications and source records attributed to Jun Chen.

At least 127 records · Page 7Linked to original sources

arXiv2025

PFEA: An LLM-based High-Level Natural Language Planning and Feedback Embodied Agent for Human-Centered AI

The rapid advancement of Large Language Models (LLMs) has marked a significant breakthrough in Artificial Intelligence (AI), ushering in a new era of Human-centered Artificial Intelligence (HAI). HAI aims to better serve human welfare and needs, thereby placing higher demands on the intelligence level of robots, particularly in aspects such as natural language interaction, complex task planning, and execution. Intelligent agents powered by LLMs have opened up new pathways for realizing HAI. However, existing LLM-based embodied agents often lack the ability to plan and execute complex natural language control tasks online. This paper explores the implementation of intelligent robotic manipulating agents based on Vision-Language Models (VLMs) in the physical world. We propose a novel embodied agent framework for robots, which comprises a human-robot voice interaction module, a vision-language agent module and an action execution module. The vision-language agent itself includes a vision-based task planner, a natural language instruction converter, and a task performance feedback evaluator. Experimental results demonstrate that our agent achieves a 28\% higher average task success rate in both simulated and real environments compared to approaches relying solely on LLM+CLIP, significantly improving the execution success rate of high-level natural language instruction tasks.

cs.RO↗

arXiv2025

DynCIM: Dynamic Curriculum for Imbalanced Multimodal Learning

Multimodal learning integrates complementary information from diverse modalities to enhance the decision-making process. However, the potential of multimodal collaboration remains under-exploited due to disparities in data quality and modality representation capabilities. To address this, we introduce DynCIM, a novel dynamic curriculum learning framework designed to quantify the inherent imbalances from both sample and modality perspectives. DynCIM employs a sample-level curriculum to dynamically assess each sample's difficulty according to prediction deviation, consistency, and stability, while a modality-level curriculum measures modality contributions from global and local. Furthermore, a gating-based dynamic fusion mechanism is introduced to adaptively adjust modality contributions, minimizing redundancy and optimizing fusion effectiveness. Extensive experiments on six multimodal benchmarking datasets, spanning both bimodal and trimodal scenarios, demonstrate that DynCIM consistently outperforms state-of-the-art methods. Our approach effectively mitigates modality and sample imbalances while enhancing adaptability and robustness in multimodal learning tasks. Our code is available at https://github.com/Raymond-Qiancx/DynCIM.

cs.CV↗

arXiv2025

Adaptive Label Correction for Robust Medical Image Segmentation with Noisy Labels

Deep learning has shown remarkable success in medical image analysis, but its reliance on large volumes of high-quality labeled data limits its applicability. While noisy labeled data are easier to obtain, directly incorporating them into training can degrade model performance. To address this challenge, we propose a Mean Teacher-based Adaptive Label Correction (ALC) self-ensemble framework for robust medical image segmentation with noisy labels. The framework leverages the Mean Teacher architecture to ensure consistent learning under noise perturbations. It includes an adaptive label refinement mechanism that dynamically captures and weights differences across multiple disturbance versions to enhance the quality of noisy labels. Additionally, a sample-level uncertainty-based label selection algorithm is introduced to prioritize high-confidence samples for network updates, mitigating the impact of noisy annotations. Consistency learning is integrated to align the predictions of the student and teacher networks, further enhancing model robustness. Extensive experiments on two public datasets demonstrate the effectiveness of the proposed framework, showing significant improvements in segmentation performance. By fully exploiting the strengths of the Mean Teacher structure, the ALC framework effectively processes noisy labels, adapts to challenging scenarios, and achieves competitive results compared to state-of-the-art methods.

cs.CV↗

arXiv2025

Fact-R1: Towards Explainable Video Misinformation Detection with Deep Reasoning

The rapid spread of multimodal misinformation on social media has raised growing concerns, while research on video misinformation detection remains limited due to the lack of large-scale, diverse datasets. Existing methods often overfit to rigid templates and lack deep reasoning over deceptive content. To address these challenges, we introduce FakeVV, a large-scale benchmark comprising over 100,000 video-text pairs with fine-grained, interpretable annotations. In addition, we further propose Fact-R1, a novel framework that integrates deep reasoning with collaborative rule-based reinforcement learning. Fact-R1 is trained through a three-stage process: (1) misinformation long-Chain-of-Thought (CoT) instruction tuning, (2) preference alignment via Direct Preference Optimization (DPO), and (3) Group Relative Policy Optimization (GRPO) using a novel verifiable reward function. This enables Fact-R1 to exhibit emergent reasoning behaviors comparable to those observed in advanced text-based reinforcement learning systems, but in the more complex multimodal misinformation setting. Our work establishes a new paradigm for misinformation detection, bridging large-scale video understanding, reasoning-guided alignment, and interpretable verification.

cs.CV↗

arXiv2025

NTIRE 2025 Challenge on Low Light Image Enhancement: Methods and Results

This paper presents a comprehensive review of the NTIRE 2025 Low-Light Image Enhancement (LLIE) Challenge, highlighting the proposed solutions and final outcomes. The objective of the challenge is to identify effective networks capable of producing brighter, clearer, and visually compelling images under diverse and challenging conditions. A remarkable total of 762 participants registered for the competition, with 28 teams ultimately submitting valid entries. This paper thoroughly evaluates the state-of-the-art advancements in LLIE, showcasing the significant progress.

cs.CV↗

arXiv2025

Vgent: Graph-based Retrieval-Reasoning-Augmented Generation For Long Video Understanding

Understanding and reasoning over long videos pose significant challenges for large video language models (LVLMs) due to the difficulty in processing intensive video tokens beyond context window and retaining long-term sequential information. Retrieval-Augmented Generation (RAG) has demonstrated effectiveness in processing long context for Large Language Models (LLMs); however, applying RAG to long video faces challenges such as disrupted temporal dependencies and inclusion of irrelevant information that can hinder accurate reasoning. To address these limitations, we propose Vgent, a novel graph-based retrieval-reasoning-augmented generation framework to enhance LVLMs for long video understanding. Our approach introduces two key innovations: (i) It represents videos by structured graphs with semantic relationships across video clips preserved to improve retrieval effectiveness. (ii) It introduces an intermediate reasoning step to mitigate the reasoning limitation of LVLMs, which leverages structured verification to reduce retrieval noise and facilitate the explicit aggregation of relevant information across clips, resulting in more accurate and context-aware responses. We comprehensively evaluate our framework with various open-source LVLMs on three long-video understanding benchmarks. Our approach yielded an overall performance improvement of $3.0\%\sim 5.4\%$ over base models on MLVU, and outperformed state-of-the-art video RAG methods by $8.6\%$. Our code is publicly available at https://xiaoqian-shen.github.io/Vgent.

cs.CV↗

arXiv2025

Frequency Domain Unlocks New Perspectives for Abdominal Medical Image Segmentation

Accurate segmentation of tumors and adjacent normal tissues in medical images is essential for surgical planning and tumor staging. Although foundation models generally perform well in segmentation tasks, they often struggle to focus on foreground areas in complex, low-contrast backgrounds, where some malignant tumors closely resemble normal organs, complicating contextual differentiation. To address these challenges, we propose the Foreground-Aware Spectrum Segmentation (FASS) framework. First, we introduce a foreground-aware module to amplify the distinction between background and the entire volume space, allowing the model to concentrate more effectively on target areas. Next, a feature-level frequency enhancement module, based on wavelet transform, extracts discriminative high-frequency features to enhance boundary recognition and detail perception. Eventually, we introduce an edge constraint module to preserve geometric continuity in segmentation boundaries. Extensive experiments on multiple medical datasets demonstrate superior performance across all metrics, validating the effectiveness of our framework, particularly in robustness under complex conditions and fine structure recognition. Our framework significantly enhances segmentation of low-contrast images, paving the way for applications in more diverse and complex medical imaging scenarios.

cs.CV↗

arXiv2025

LSZone: A Lightweight Spatial Information Modeling Architecture for Real-time In-car Multi-zone Speech Separation

In-car multi-zone speech separation, which captures voices from different speech zones, plays a crucial role in human-vehicle interaction. Although previous SpatialNet has achieved notable results, its high computational cost still hinders real-time applications in vehicles. To this end, this paper proposes LSZone, a lightweight spatial information modeling architecture for real-time in-car multi-zone speech separation. We design a spatial information extraction-compression (SpaIEC) module that combines Mel spectrogram and Interaural Phase Difference (IPD) to reduce computational burden while maintaining performance. Additionally, to efficiently model spatial information, we introduce an extremely lightweight Conv-GRU crossband-narrowband processing (CNP) module. Experimental results demonstrate that LSZone, with a complexity of 0.56G MACs and a real-time factor (RTF) of 0.37, delivers impressive performance in complex noise and multi-speaker scenarios.

cs.SD↗

arXiv2025

RainSeer: Fine-Grained Rainfall Reconstruction via Physics-Guided Modeling

Reconstructing high-resolution rainfall fields is essential for flood forecasting, hydrological modeling, and climate analysis. However, existing spatial interpolation methods-whether based on automatic weather station (AWS) measurements or enhanced with satellite/radar observations often over-smooth critical structures, failing to capture sharp transitions and localized extremes. We introduce RainSeer, a structure-aware reconstruction framework that reinterprets radar reflectivity as a physically grounded structural prior-capturing when, where, and how rain develops. This shift, however, introduces two fundamental challenges: (i) translating high-resolution volumetric radar fields into sparse point-wise rainfall observations, and (ii) bridging the physical disconnect between aloft hydro-meteors and ground-level precipitation. RainSeer addresses these through a physics-informed two-stage architecture: a Structure-to-Point Mapper performs spatial alignment by projecting mesoscale radar structures into localized ground-level rainfall, through a bidirectional mapping, and a Geo-Aware Rain Decoder captures the semantic transformation of hydro-meteors through descent, melting, and evaporation via a causal spatiotemporal attention mechanism. We evaluate RainSeer on two public datasets-RAIN-F (Korea, 2017-2019) and MeteoNet (France, 2016-2018)-and observe consistent improvements over state-of-the-art baselines, reducing MAE by over 13.31% and significantly enhancing structural fidelity in reconstructed rainfall fields.

cs.LG↗

arXiv2025

A Hierarchical Geometry-guided Transformer for Histological Subtyping of Primary Liver Cancer

Primary liver malignancies are widely recognized as the most heterogeneous and prognostically diverse cancers of the digestive system. Among these, hepatocellular carcinoma (HCC) and intrahepatic cholangiocarcinoma (ICC) emerge as the two principal histological subtypes, demonstrating significantly greater complexity in tissue morphology and cellular architecture than other common tumors. The intricate representation of features in Whole Slide Images (WSIs) encompasses abundant crucial information for liver cancer histological subtyping, regarding hierarchical pyramid structure, tumor microenvironment (TME), and geometric representation. However, recent approaches have not adequately exploited these indispensable effective descriptors, resulting in a limited understanding of histological representation and suboptimal subtyping performance. To mitigate these limitations, ARGUS is proposed to advance histological subtyping in liver cancer by capturing the macro-meso-micro hierarchical information within the TME. Specifically, we first construct a micro-geometry feature to represent fine-grained cell-level pattern via a geometric structure across nuclei, thereby providing a more refined and precise perspective for delineating pathological images. Then, a Hierarchical Field-of-Views (FoVs) Alignment module is designed to model macro- and meso-level hierarchical interactions inherent in WSIs. Finally, the augmented micro-geometry and FoVs features are fused into a joint representation via present Geometry Prior Guided Fusion strategy for modeling holistic phenotype interactions. Extensive experiments on public and private cohorts demonstrate that our ARGUS achieves state-of-the-art (SOTA) performance in histological subtyping of liver cancer, which provide an effective diagnostic tool for primary liver malignancies in clinical practice.

cs.CV↗

arXiv2025

Electric-field control of pure spin photocurrent in germanene

The electrical control of pure spin current remains a central challenge in spintronics, particularly in time-reversal symmetric systems composed of nonmagnetic elements, where spin and electric fields interact only indirectly. In this work, we develop a theoretical framework for electrically tuning pure spin photocurrent in two-dimensional materials with time-reversal symmetry via a gate electric field. Through theoretical analysis, we demonstrate that in systems with spin-orbit coupling and in-plane mirror symmetry, an out-of-plane electric field induces spin splitting and reversal in the band structure near the Fermi energy, enabling magnitude control and direction reversal of the pure spin photocurrent. To validate this mechanism, we perform first-principles calculations on germanene, an experimentally realized two-dimensional material. Beyond amplitude modulation, we reveal that reversing the direction of the applied electric field leads to a corresponding reversal of the pure spin photocurrent. Furthermore, we show that the pure spin photocurrent can be tuned by varying the photon energy and the incident angle of light, providing additional degrees of control over spin transport. These findings establish a robust strategy for electric-field-controlled pure spin transport in two-dimensional materials, offering new possibilities for the development of optospintronic devices.

cond-mat.mes-hall↗

arXiv2025

Neural Catalog: Scaling Species Recognition with Catalog of Life-Augmented Generation

Open-vocabulary species recognition is a major challenge in computer vision, particularly in ornithology, where new taxa are continually discovered. While benchmarks like CUB-200-2011 and Birdsnap have advanced fine-grained recognition under closed vocabularies, they fall short of real-world conditions. We show that current systems suffer a performance drop of over 30\% in realistic open-vocabulary settings with thousands of candidate species, largely due to an increased number of visually similar and semantically ambiguous distractors. To address this, we propose Visual Re-ranking Retrieval-Augmented Generation (VR-RAG), a novel framework that links structured encyclopedic knowledge with recognition. We distill Wikipedia articles for 11,202 bird species into concise, discriminative summaries and retrieve candidates from these summaries. Unlike prior text-only approaches, VR-RAG incorporates visual information during retrieval, ensuring final predictions are both textually relevant and visually consistent with the query image. Extensive experiments across five bird classification benchmarks and two additional domains show that VR-RAG improves the average performance of the state-of-the-art Qwen2.5-VL model by 18.0%.

cs.CV↗

arXiv2025

Distributed Multi-Robot Multi-Target Simultaneous Search and Tracking in an Unknown Non-convex Environment

In unknown non-convex environments, such as indoor and underground spaces, deploying a fleet of robots to explore the surroundings while simultaneously searching for and tracking targets of interest to maintain high-precision data collection represents a fundamental challenge that urgently requires resolution in applications such as environmental monitoring and rescue operations. Current research has made significant progress in addressing environmental exploration, information search, and target tracking problems, but has yet to establish a framework for simultaneously optimizing these tasks in complex environments. In this paper, we propose a novel motion planning algorithm framework that integrates three control strategies: a frontier-based exploration strategy, a guaranteed coverage strategy based on Lloyd's algorithm, and a sensor-based multi-target tracking strategy. By incorporating these three strategies, the proposed algorithm balances coverage search and high-precision active tracking during exploration. Our approach is validated through a series of MATLAB simulations, demonstrating validity and superiority over standard approaches.

cs.RO↗

arXiv2025

Learning Probabilistic Obstacle Spaces from Data-driven Uncertainty using Neural Networks

Identifying the obstacle space is crucial for path planning. However, generating an accurate obstacle space remains a significant challenge due to various sources of uncertainty, including motion, behavior, and perception limitations. Even though an autonomous system can operate with an inaccurate obstacle space by being over-conservative and using redundant sensors, a more accurate obstacle space generator can reduce both path planning costs and hardware costs. Existing generation methods that generate high-quality output are all computationally expensive. Traditional methods, such as filtering, sensor fusion and data-driven estimators, face significant computational challenges or require large amounts of data, which limits their applicability in realistic scenarios. In this paper, we propose leveraging neural networks, commonly used in imitation learning, to mimic expert methods for modeling uncertainty and generating confidence regions for obstacle positions, which we refer to as the probabilistic obstacle space. The network is trained using a multi-label, supervised learning approach. We adopt a fine-tuned convex approximation method as the expert to construct training datasets. After training, given only a small number of samples, the neural network can accurately replicate the probabilistic obstacle space while achieving substantially faster generation speed. Moreover, the resulting obstacle space is convex, making it more convenient for subsequent path planning.

cs.RO↗

arXiv2025

AEFS: Adaptive Early Feature Selection for Deep Recommender Systems

Feature selection has emerged as a crucial technique in refining recommender systems. Recent advancements leveraging Automated Machine Learning (AutoML) has drawn significant attention, particularly in two main categories: early feature selection and late feature selection, differentiated by whether the selection occurs before or after the embedding layer. The early feature selection selects a fixed subset of features and retrains the model, while the late feature selection, known as adaptive feature selection, dynamically adjusts feature choices for each data instance, recognizing the variability in feature significance. Although adaptive feature selection has shown remarkable improvements in performance, its main drawback lies in its post-embedding layer feature selection. This process often becomes cumbersome and inefficient in large-scale recommender systems with billions of ID-type features, leading to a highly sparse and parameter-heavy embedding layer. To overcome this, we introduce Adaptive Early Feature Selection (AEFS), a very simple method that not only adaptively selects informative features for each instance, but also significantly reduces the activated parameters of the embedding layer. AEFS employs a dual-model architecture, encompassing an auxiliary model dedicated to feature selection and a main model responsible for prediction. To ensure effective alignment between these two models, we incorporate two collaborative training loss constraints. Our extensive experiments on three benchmark datasets validate the efficiency and effectiveness of our approach. Notably, AEFS matches the performance of current state-of-theart Adaptive Late Feature Selection methods while achieving a significant reduction of 37. 5% in the activated parameters of the embedding layer. AEFS is open-source at https://github. com/fly-dragon211/AEFS .

cs.IR↗

arXiv2025

WikiAutoGen: Towards Multi-Modal Wikipedia-Style Article Generation

Knowledge discovery and collection are intelligence-intensive tasks that traditionally require significant human effort to ensure high-quality outputs. Recent research has explored multi-agent frameworks for automating Wikipedia-style article generation by retrieving and synthesizing information from the internet. However, these methods primarily focus on text-only generation, overlooking the importance of multimodal content in enhancing informativeness and engagement. In this work, we introduce WikiAutoGen, a novel system for automated multimodal Wikipedia-style article generation. Unlike prior approaches, WikiAutoGen retrieves and integrates relevant images alongside text, enriching both the depth and visual appeal of generated content. To further improve factual accuracy and comprehensiveness, we propose a multi-perspective self-reflection mechanism, which critically assesses retrieved content from diverse viewpoints to enhance reliability, breadth, and coherence, etc. Additionally, we introduce WikiSeek, a benchmark comprising Wikipedia articles with topics paired with both textual and image-based representations, designed to evaluate multimodal knowledge generation on more challenging topics. Experimental results show that WikiAutoGen outperforms previous methods by 8%-29% on our WikiSeek benchmark, producing more accurate, coherent, and visually enriched Wikipedia-style articles. Our code and examples are available at https://wikiautogen.github.io/

cs.CV↗

arXiv2025

The Role of Far-side Magnetic Structures in Modeling 2024 Solar Eclipse

The corona is a crucial region that connects the solar surface to the solar wind and serves as the primary site of solar activity. The 2024 total solar eclipse (TSE) provides a unique opportunity to investigate the large-scale coronal structure. Combined with TSE observations, we study the impact of the magnetic structure of the far-side active region, located in the eastern hemisphere of the Sun that has not yet rotated into the Earth Field-of-View (FoV), on a global Magnetohydrodynamic (MHD) simulation. To address the limitation of single-view measurements in the routine synoptic map, we correct the magnetic field in the far-side region by incorporating full-disk magnetograms measured several days after the TSE, allowing us to capture the temporal evolution of the photospheric magnetic field in near real-time. Simulation results demonstrate that the local magnetic field in the far-side active region can significantly influence the global coronal structure by altering the position of the heliospheric current sheet (HCS), and further affect the global distribution of plasma parameters, even in polar regions. A comparison of the simulation results with white-light (WL) TSE + LASCO C2 observations and in situ measurements by the Parker Solar Probe (PSP) reveals that the composite synoptic map improves the accuracy of coronal modeling. This work provides robust support for advancing our understanding of coronal evolution, as well as deepens the link between the photosphere and large-scale coronal structure. Furthermore, it establishes a theoretical foundation for the future development of multi-view, stereoscopic measurements of the photospheric magnetic field.

astro-ph.SR↗

arXiv2025

Quantifying non-equilibrium pressure-gradient turbulent boundary layers through a symmetry-based framework

This study establishes a symmetry-based framework to quantify non-equilibrium processes in complex pressure gradient (PG) turbulent boundary layers (TBLs), using a Lie-group-informed dilation-symmetry-breaking formalism. We derive a universal multilayer defect scaling law for the evolution of total shear stress (TSS). The law shows that gradually varying adverse pressure gradients (APGs) break the dilation symmetry in the two-layer defect scaling of equilibrium TSS, leading to three-layer TSS structures. For abrupt PG transitions, we identify boundary-layer decoupling into: 1) an equilibrium internal boundary layer, and 2) a history-dependent outer flow, arising from disparate adaptation timescales. The framework introduces a unified velocity scale mapping non-equilibrium TSS to canonical zero-PG scaling. Validation spans different aerodynamic systems, including developing APG on airfoils, APG-to-favorable-PG transition on a Gaussian bump, and favorable PG to rapidly amplifying APG in a converging-diverging channel. The work enables improved prediction of non-equilibrium PG TBL behavior through unified characterization of stress evolution dynamics, providing new physics-based parameterizations that could promote machine learning of complex wall-bounded turbulent flows.

physics.flu-dyn↗