arXiv subjects

Lei Fan

Publications and source records attributed to Lei Fan.

At least 55 records · Page 3Linked to original sources

arXiv2025

Flexible Concept Bottleneck Model

Concept bottleneck models (CBMs) improve neural network interpretability by introducing an intermediate layer that maps human-understandable concepts to predictions. Recent work has explored the use of vision-language models (VLMs) to automate concept selection and annotation. However, existing VLM-based CBMs typically require full model retraining when new concepts are involved, which limits their adaptability and flexibility in real-world scenarios, especially considering the rapid evolution of vision-language foundation models. To address these issues, we propose Flexible Concept Bottleneck Model (FCBM), which supports dynamic concept adaptation, including complete replacement of the original concept set. Specifically, we design a hypernetwork that generates prediction weights based on concept embeddings, allowing seamless integration of new concepts without retraining the entire model. In addition, we introduce a modified sparsemax module with a learnable temperature parameter that dynamically selects the most relevant concepts, enabling the model to focus on the most informative features. Extensive experiments on five public benchmarks demonstrate that our method achieves accuracy comparable to state-of-the-art baselines with a similar number of effective concepts. Moreover, the model generalizes well to unseen concepts with just a single epoch of fine-tuning, demonstrating its strong adaptability and flexibility.

cs.CV↗

arXiv2025

Strategic Decision-Making Under Uncertainty through Bi-Level Game Theory and Distributionally Robust Optimization

In strategic scenarios where decision-makers operate at different hierarchical levels, traditional optimization methods are often inadequate for handling uncertainties from incomplete information or unpredictable external factors. To fill this gap, we introduce a mathematical framework that integrates bi-level game theory with distributionally robust optimization (DRO), particularly suited for complex network systems. Our approach leverages the hierarchical structure of bi-level games to model leader-follower interactions while incorporating distributional robustness to guard against worst-case probability distributions. To ensure computational tractability, the Karush-Kuhn-Tucker (KKT) conditions are used to transform the bi-level challenge into a more manageable single-level model, and the infinite-dimensional DRO problem is reformulated into a finite equivalent. We propose a generalized algorithm to solve this integrated model. Simulation results validate our framework's efficacy, demonstrating that under high uncertainty, the proposed model achieves up to a 22\% cost reduction compared to traditional stochastic methods while maintaining a service level of over 90\%. This highlights its potential to significantly improve decision quality and robustness in networked systems such as transportation and communication networks.

eess.SY↗

arXiv2025

SwinMamba: A hybrid local-global mamba framework for enhancing semantic segmentation of remotely sensed images

Semantic segmentation of remote sensing imagery is a fundamental task in computer vision, supporting a wide range of applications such as land use classification, urban planning, and environmental monitoring. However, this task is often challenged by the high spatial resolution, complex scene structures, and diverse object scales present in remote sensing data. To address these challenges, various deep learning architectures have been proposed, including convolutional neural networks, Vision Transformers, and the recently introduced Vision Mamba. Vision Mamba features a global receptive field and low computational complexity, demonstrating both efficiency and effectiveness in image segmentation. However, its reliance on global scanning tends to overlook critical local features, such as textures and edges, which are essential for achieving accurate segmentation in remote sensing contexts. To tackle this limitation, we propose SwinMamba, a novel framework inspired by the Swin Transformer. SwinMamba integrates localized Mamba-style scanning within shifted windows with a global receptive field, to enhance the model's perception of both local and global features. Specifically, the first two stages of SwinMamba perform local scanning to capture fine-grained details, while its subsequent two stages leverage global scanning to fuse broader contextual information. In our model, the use of overlapping shifted windows enhances inter-region information exchange, facilitating more robust feature integration across the entire image. Extensive experiments on the LoveDA and ISPRS Potsdam datasets demonstrate that SwinMamba outperforms state-of-the-art methods, underscoring its effectiveness and potential as a superior solution for semantic segmentation of remotely sensed imagery.

cs.CV↗

arXiv2025

SOLAR: Switchable Output Layer for Accuracy and Robustness in Once-for-All Training

Once-for-All (OFA) training enables a single super-net to generate multiple sub-nets tailored to diverse deployment scenarios, supporting flexible trade-offs among accuracy, robustness, and model-size without retraining. However, as the number of supported sub-nets increases, excessive parameter sharing in the backbone limits representational capacity, leading to degraded calibration and reduced overall performance. To address this, we propose SOLAR (Switchable Output Layer for Accuracy and Robustness in Once-for-All Training), a simple yet effective technique that assigns each sub-net a separate classification head. By decoupling the logit learning process across sub-nets, the Switchable Output Layer (SOL) reduces representational interference and improves optimization, without altering the shared backbone. We evaluate SOLAR on five datasets (SVHN, CIFAR-10, STL-10, CIFAR-100, and TinyImageNet) using four super-net backbones (ResNet-34, WideResNet-16-8, WideResNet-40-2, and MobileNetV2) for two OFA training frameworks (OATS and SNNs). Experiments show that SOLAR outperforms the baseline methods: compared to OATS, it improves accuracy of sub-nets up to 1.26 %, 4.71 %, 1.67 %, and 1.76 %, and robustness up to 9.01 %, 7.71 %, 2.72 %, and 1.26 % on SVHN, CIFAR-10, STL-10, and CIFAR-100, respectively. Compared to SNNs, it improves TinyImageNet accuracy by up to 2.93 %, 2.34 %, and 1.35 % using ResNet-34, WideResNet-16-8, and MobileNetV2 backbones (with 8 sub-nets), respectively.

cs.LG↗

arXiv2025

Self-Rewarding Rubric-Based Reinforcement Learning for Open-Ended Reasoning

Open-ended evaluation is essential for deploying large language models in real-world settings. In studying HealthBench, we observe that using the model itself as a grader and generating rubric-based reward signals substantially improves reasoning performance. Remarkably, the trained model also becomes a stronger grader. Motivated by this, we introduce Self-Rewarding Rubric-Based Reinforcement Learning for Open-Ended Reasoning, a lightweight framework that enables faster and more resource-efficient training while surpassing baselines. Remarkably, on Qwen3-32B, training with just the 4000-sample HealthBench Easy subset is sufficient to obtain a model that exceeds GPT-5 on HealthBench Hard. Incorporating a small amount of teacher-graded data further enhances performance for less capable models.

cs.CL↗

arXiv2025

JUNO 20-inch PMT and electronics system characterization using large pulses of PMT dark counts at the Pan-Asia testing platform

The main goal of the JUNO experiment is to determine the neutrino mass ordering with a 20kt liquid-scintillator detector. The 20-inch PMT and its 1F3 (one for three) electronics are crucial to realize the excellent energy resolution of at least 3% at 1MeV. The knowledge on the PMT and 1F3 electronics response is critical for detector performance understanding. A study of the JUNO 20-inch PMT and 1F3 electronics system characterization is presented using large pulses of PMT dark count at the Pan-Asia testing platform in China. Thanks to its broad amplitude range and high rate, the large pulse signals are also used to investigate the PMT after pulse response.

physics.ins-det↗

arXiv2025

Efficient Entanglement Routing for Satellite-Aerial-Terrestrial Quantum Networks

In the era of 6G and beyond, space-aerial-terrestrial quantum networks (SATQNs) are shaping the future of the global-scale quantum Internet. This paper investigates the collaboration among satellite, aerial, and terrestrial quantum networks to efficiently transmit high-fidelity quantum entanglements over long distances. We begin with a comprehensive overview of existing satellite-, aerial-, and terrestrial-based quantum networks. Subsequently, we address the entanglement routing problem with the objective of maximizing quantum network throughput by jointly optimizing path selection and entanglement generation rates (PS-EGR). Given that the original problem is formulated as a mixed-integer linear programming (MILP) problem, which is inherently intractable, we propose a Benders' decomposition (BD)-based algorithm to solve the problem efficiently. Numerical results validate the effectiveness of the proposed PS-EGR scheme, offering valuable insights into various optimizable factors within the system. Finally, we discuss the current challenges and propose promising avenues for future research in SATQNs.

quant-ph↗

arXiv2025

Learning to Align, Aligning to Learn: A Unified Approach for Self-Optimized Alignment

Alignment methodologies have emerged as a critical pathway for enhancing language model alignment capabilities. While SFT (supervised fine-tuning) accelerates convergence through direct token-level loss intervention, its efficacy is constrained by offline policy trajectory. In contrast, RL(reinforcement learning) facilitates exploratory policy optimization, but suffers from low sample efficiency and stringent dependency on high-quality base models. To address these dual challenges, we propose GRAO (Group Relative Alignment Optimization), a unified framework that synergizes the respective strengths of SFT and RL through three key innovations: 1) A multi-sample generation strategy enabling comparative quality assessment via reward feedback; 2) A novel Group Direct Alignment Loss formulation leveraging intra-group relative advantage weighting; 3) Reference-aware parameter updates guided by pairwise preference dynamics. Our theoretical analysis establishes GRAO's convergence guarantees and sample efficiency advantages over conventional approaches. Comprehensive evaluations across complex human alignment tasks demonstrate GRAO's superior performance, achieving 57.70\%,17.65\% 7.95\% and 5.18\% relative improvements over SFT, DPO, PPO and GRPO baselines respectively. This work provides both a theoretically grounded alignment framework and empirical evidence for efficient capability evolution in language models.

cs.LG↗

arXiv2025

Static and Plugged: Make Embodied Evaluation Simple

Embodied intelligence is advancing rapidly, driving the need for efficient evaluation. Current benchmarks typically rely on interactive simulated environments or real-world setups, which are costly, fragmented, and hard to scale. To address this, we introduce StaticEmbodiedBench, a plug-and-play benchmark that enables unified evaluation using static scene representations. Covering 42 diverse scenarios and 8 core dimensions, it supports scalable and comprehensive assessment through a simple interface. Furthermore, we evaluate 19 Vision-Language Models (VLMs) and 11 Vision-Language-Action models (VLAs), establishing the first unified static leaderboard for Embodied intelligence. Moreover, we release a subset of 200 samples from our benchmark to accelerate the development of embodied intelligence.

cs.CV↗

arXiv2025

Quantum Hamiltonian Descent based Augmented Lagrangian Method for Constrained Nonconvex Nonlinear Optimization

Nonlinear programming (NLP) plays a critical role in domains such as power energy systems, chemical engineering, communication networks, and financial engineering. However, solving large-scale, nonconvex NLP problems remains a significant challenge due to the complexity of the solution landscape and the presence of nonlinear nonconvex constraints. In this paper, we develop a Quantum Hamiltonian Descent based Augmented Lagrange Method (QHD-ALM) framework to address largescale, constrained nonconvex NLP problems. The augmented Lagrange method (ALM) can convert a constrained NLP to an unconstrained NLP, which can be solved by using Quantum Hamiltonian Descent (QHD). To run the QHD on a classical machine, we propose to use the Simulated Bifurcation algorithm as the engine to simulate the dynamic process. We apply our algorithm to a Power-to-Hydrogen System, and the simulation results verify the effectiveness of our algorithm.

math.OC↗

arXiv2025

Low-loss terahertz negative curvature suspended-core fiber

Inspired by the design concept of negative curvature hollow-core fibers, this paper presents an innovative negative curvature suspended-core THz fiber. Compared to traditional suspended-core fibers, all structural units of this fiber are designed with circular boundaries, effectively avoiding the issues of insufficient mechanical strength and manufacturing difficulties associated with the wide and thin rectangular support arms in traditional structures. The numerical simulation using the full-vector finite element method shows that the optical fiber loss is as low as 0.02cm-1 in 0.66-1.09THz, and the low loss bandwidth is 0.43THz. In addition, near-zero flat dispersion of -0.08-0.74 ps/THz/cm can be achieved. The fiber exhibits excellent characteristics of low loss, wide bandwidth, and low dispersion, theoretically opening a new research path for the design of low-loss THz fibers.

physics.optics↗

arXiv2025

DiTalker: A Unified DiT-based Framework for High-Quality and Speaking Styles Controllable Portrait Animation

Portrait animation aims to synthesize talking videos from a static reference face, conditioned on audio and style frame cues (e.g., emotion and head poses), while ensuring precise lip synchronization and faithful reproduction of speaking styles. Existing diffusion-based portrait animation methods primarily focus on lip synchronization or static emotion transformation, often overlooking dynamic styles such as head movements. Moreover, most of these methods rely on a dual U-Net architecture, which preserves identity consistency but incurs additional computational overhead. To this end, we propose DiTalker, a unified DiT-based framework for speaking style-controllable portrait animation. We design a Style-Emotion Encoding Module that employs two separate branches: a style branch extracting identity-specific style information (e.g., head poses and movements), and an emotion branch extracting identity-agnostic emotion features. We further introduce an Audio-Style Fusion Module that decouples audio and speaking styles via two parallel cross-attention layers, using these features to guide the animation process. To enhance the quality of results, we adopt and modify two optimization constraints: one to improve lip synchronization and the other to preserve fine-grained identity and background details. Extensive experiments demonstrate the superiority of DiTalker in terms of lip synchronization and speaking style controllability. Project Page: https://thenameishope.github.io/DiTalker/

cs.CV↗

arXiv2025

SAGE: A Visual Language Model for Anomaly Detection via Fact Enhancement and Entropy-aware Alignment

While Vision-Language Models (VLMs) have shown promising progress in general multimodal tasks, they often struggle in industrial anomaly detection and reasoning, particularly in delivering interpretable explanations and generalizing to unseen categories. This limitation stems from the inherently domain-specific nature of anomaly detection, which hinders the applicability of existing VLMs in industrial scenarios that require precise, structured, and context-aware analysis. To address these challenges, we propose SAGE, a VLM-based framework that enhances anomaly reasoning through Self-Guided Fact Enhancement (SFE) and Entropy-aware Direct Preference Optimization (E-DPO). SFE integrates domain-specific knowledge into visual reasoning via fact extraction and fusion, while E-DPO aligns model outputs with expert preferences using entropy-aware optimization. Additionally, we introduce AD-PL, a preference-optimized dataset tailored for industrial anomaly reasoning, consisting of 28,415 question-answering instances with expert-ranked responses. To evaluate anomaly reasoning models, we develop Multiscale Logical Evaluation (MLE), a quantitative framework analyzing model logic and consistency. SAGE demonstrates superior performance on industrial anomaly datasets under zero-shot and one-shot settings. The code, model and dataset are available at https://github.com/amoreZgx1n/SAGE.

cs.CL↗

arXiv2025

Salvaging the Overlooked: Leveraging Class-Aware Contrastive Learning for Multi-Class Anomaly Detection

For anomaly detection (AD), early approaches often train separate models for individual classes, yielding high performance but posing challenges in scalability and resource management. Recent efforts have shifted toward training a single model capable of handling multiple classes. However, directly extending early AD methods to multi-class settings often results in degraded performance. In this paper, we investigate this performance degradation observed in reconstruction-based methods, identifying the key issue: inter-class confusion. This confusion emerges when a model trained in multi-class scenarios incorrectly reconstructs samples from one class as those of another, thereby exacerbating reconstruction errors. To this end, we propose a simple yet effective modification, called class-aware contrastive learning (CCL). By explicitly leveraging raw object category information (\eg carpet or wood) as supervised signals, we introduce local CL to refine multiscale dense features, and global CL to obtain more compact feature representations of normal patterns, thereby effectively adapting the models to multi-class settings. Experiments across five datasets validate the effectiveness of our approach, demonstrating significant improvements and superior performance compared to state-of-the-art methods. Notably, ablation studies indicate that pseudo-class labels can achieve comparable performance.

cs.CV↗

arXiv2025

A review of advancements in low-light image enhancement using deep learning

In low-light environments, the performance of computer vision algorithms often deteriorates significantly, adversely affecting key vision tasks such as segmentation, detection, and classification. With the rapid advancement of deep learning, its application to low-light image processing has attracted widespread attention and seen significant progress in recent years. However, there remains a lack of comprehensive surveys that systematically examine how recent deep-learning-based low-light image enhancement methods function and evaluate their effectiveness in enhancing downstream vision tasks. To address this gap, this review provides detailed elaboration on how various recent approaches (from 2020) operate and their enhancement mechanisms, supplemented with clear illustrations. It also investigates the impact of different enhancement techniques on subsequent vision tasks, critically analyzing their strengths and limitations. Our review found that image enhancement improved the performance of downstream vision tasks to varying degrees. Although supervised methods often produced images with high perceptual quality, they typically produced modest improvements in vision tasks. In contrast, zero-shot learning, despite achieving lower scores in image quality metrics, showed consistently boosted performance across various vision tasks. These suggest a disconnect between image quality metrics and those evaluating vision task performance. Additionally, unsupervised domain adaptation techniques demonstrated significant gains in segmentation tasks, highlighting their potential in practical low-light scenarios where labelled data is scarce. Observed limitations of existing studies are analyzed, and directions for future research are proposed. This review serves as a useful reference for determining low-light image enhancement techniques and optimizing vision task performance in low-light conditions.

cs.CV↗

arXiv2025

DH-FaceVid-1K: A Large-Scale High-Quality Dataset for Face Video Generation

Human-centric generative models are becoming increasingly popular, giving rise to various innovative tools and applications, such as talking face videos conditioned on text or audio prompts. The core of these capabilities lies in powerful pre-trained foundation models, trained on large-scale, high-quality datasets. However, many advanced methods rely on in-house data subject to various constraints, and other current studies fail to generate high-resolution face videos, which is mainly attributed to the significant lack of large-scale, high-quality face video datasets. In this paper, we introduce a human face video dataset, \textbf{DH-FaceVid-1K}. Our collection spans 1,200 hours in total, encompassing 270,043 video clips from over 20,000 individuals. Each sample includes corresponding speech audio, facial keypoints, and text annotations. Compared to other publicly available datasets, ours distinguishes itself through its multi-ethnic coverage and high-quality, comprehensive individual attributes. We establish multiple face video generation models supporting tasks such as text-to-video and image-to-video generation. In addition, we develop comprehensive benchmarks to validate the scaling law when using different proportions of proposed dataset. Our primary aim is to contribute a face video dataset, particularly addressing the underrepresentation of Asian faces in existing curated datasets and thereby enriching the global spectrum of face-centric data and mitigating demographic biases. \textbf{Project Page:} https://luna-ai-lab.github.io/DH-FaceVid-1K/

cs.CV↗

arXiv2025

LLaPa: A Vision-Language Model Framework for Counterfactual-Aware Procedural Planning

While large language models (LLMs) have advanced procedural planning for embodied AI systems through strong reasoning abilities, the integration of multimodal inputs and counterfactual reasoning remains underexplored. To tackle these challenges, we introduce LLaPa, a vision-language model framework designed for multimodal procedural planning. LLaPa generates executable action sequences from textual task descriptions and visual environmental images using vision-language models (VLMs). Furthermore, we enhance LLaPa with two auxiliary modules to improve procedural planning. The first module, the Task-Environment Reranker (TER), leverages task-oriented segmentation to create a task-sensitive feature space, aligning textual descriptions with visual environments and emphasizing critical regions for procedural execution. The second module, the Counterfactual Activities Retriever (CAR), identifies and emphasizes potential counterfactual conditions, enhancing the model's reasoning capability in counterfactual scenarios. Extensive experiments on ActPlan-1K and ALFRED benchmarks demonstrate that LLaPa generates higher-quality plans with superior LCS and correctness, outperforming advanced models. The code and models are available https://github.com/sunshibo1234/LLaPa.

cs.CL↗

arXiv2025

Spatially Gene Expression Prediction using Dual-Scale Contrastive Learning

Spatial transcriptomics (ST) provides crucial insights into tissue micro-environments, but is limited to its high cost and complexity. As an alternative, predicting gene expression from pathology whole slide images (WSI) is gaining increasing attention. However, existing methods typically rely on single patches or a single pathology modality, neglecting the complex spatial and molecular interactions between target and neighboring information (e.g., gene co-expression). This leads to a failure in establishing connections among adjacent regions and capturing intricate cross-modal relationships. To address these issues, we propose NH2ST, a framework that integrates spatial context and both pathology and gene modalities for gene expression prediction. Our model comprises a query branch and a neighbor branch to process paired target patch and gene data and their neighboring regions, where cross-attention and contrastive learning are employed to capture intrinsic associations and ensure alignments between pathology and gene expression. Extensive experiments on six datasets demonstrate that our model consistently outperforms existing methods, achieving over 20% in PCC metrics. Codes are available at https://github.com/MCPathology/NH2ST

cs.CV↗