检索结果-内蒙古大学图书馆

Out-of-Distribution Detection by Principal Component Correspondence

学校读者我要写书评

暂无评论

Out-of-Distribution Detection by Principal Component Corresp...

IEEE International Conference on Multimedia and Expo (ICME)

作者： Xiaoyuan Guan Zhiyong Gan Ling Deng Wei Shi Jiankang Chen Shenshen Bu Chunliang Zhao Jianfang Hu Yuren Zhou Wei-Shi Zheng Ruixuan Wang School of Computer Science and Engineering Sun Yat-Sen University Guangzhou China Key Laboratory of Machine Intelligence and Advanced Computing MOE Guangzhou China Network Business Group China Unicom Guangzhou China Department of Network Intelligence Pengcheng Laboratory Shenzhen China School of Data and Science Qingdao University of Science and Technology Qingdao China School of Software Engineering Sun Yat-Sen University Zhuhai China

ISBN: (数字)9798350390155

ISBN: (纸本)9798350390162

Out-of-distribution (OOD) detection is vital for the safe application of intelligent systems in real-world scenarios. This paper proposes an enhancement to OOD detection by leveraging the consistency in cognition between two models, both pretrained on in-distribution (ID) data. Specifically, for a given test sample, we first apply Principal Component Analysis (PCA)-based projection on the feature vectors from each model. These obtained feature vectors (with correlation between dimensions decoupled by PCA projection) are then aligned using a multiple linear mapping, which is fitted using the least squares method on the training data. We hypothesize that the regression error for OOD data will be larger than that for ID data, making it a useful metric for OOD detection. Our experimental results demonstrate the effectiveness of this method. When combined with existing robust baselines, our approach achieves state-of-the-art performance in OOD detection.

关键词： Training Measurement Analytical models Training data Vectors Data models Intelligent systems Facial features Principal component analysis Least mean squares methods

Network Embedding with Topology-Aware Textual Representations 18th

学校读者我要写书评

暂无评论

Network Embedding with Topology-Aware Textual Representation...

18th Pacific Rim International Conference on Artificial intelligence, PRICAI 2021

作者： Chen, Jiaxing Xu, Zenan Su, Qinliang School of Computer Science and Engineering Sun Yat-sen University Guangzhou China Guangdong Key Laboratory of Big Data Analysis and Processing Guangzhou China Key Laboratory of Machine Intelligence and Advanced Computing Ministry of Education Guangzhou China

ISBN: (纸本)9783030891879

Textual network embedding aims to learn meaningful low-dimensional representations for vertices with the consideration of the associated texts. When learning the representations for texts in network embedding, existing methods mostly only exploit information from neighboring texts (i.e., contexts), while rarely taking advantages of the valuable network topological (structural) information. To bridge the gap, in this paper, a model based on adaptive-filter convolutional neural networks (CNN) is developed, in which the filters are adapted to local network topologies, rather than clamped to fixed values as in traditional CNNs. The dependency enables the learned text representations to be aware of local network topologies. It is shown that the proposed topology-aware representations can be viewed as a complement to existing context-aware ones. When the two are used together, experimental results on three real-world benchmarks demonstrate that significant performance improvements on the tasks of link prediction and vertex classification. © 2021, Springer Nature Switzerland AG.

关键词： Network topology

Self-Attention Prediction Correction with Channel Suppression for Weakly-Supervised Semantic Segmentation

学校读者我要写书评

暂无评论

Self-Attention Prediction Correction with Channel Suppressio...

IEEE International Conference on Multimedia and Expo (ICME)

作者： Guoying Sun Meng Yang School of Computer Science and Engineering Sun Yat-sen University Guangzhou China State Key Laboratory of Integrated Services Networks Xidian University Key Laboratory of Machine Intelligence and Advanced Computing Sun Yat-sen University Ministry of Education China

Single-stage weakly-supervised semantic segmentation (WSSS) with image-level labels has become a new research hotspot in the community for its lower cost and higher training efficiency. However, the pseudo label of WSSS generally suffers from somewhat noise, which limits the segmentation performance. In this paper, to explore the integral foreground activation, we propose the Channel Suppression (CS) module for preventing only activating the most discriminative regions, thereby improving the initial pseudo labels. To rectify the in-correct prediction, we explore the Self-Attention Prediction Correction (SAPC) module, which adaptively generates the category-wise prediction rectification weights. After extensive experiments, the proposed efficient single-stage framework achieves excellent performance with 67.6% mIoU and 39.9% mIoU on PASCAL VOC 2012 and MS COCO 2014 datasets, significantly exceeding several recent single-stage methods.

关键词：

A Lane Detection Method for Advancing Sustainable Energy Internet

学校读者我要写书评

暂无评论

A Lane Detection Method for Advancing Sustainable Energy Int...

IEEE Conference on Energy Internet and Energy System Integration (EI2)

作者： Xiting Peng Yang Yang Xiaoyu Zhang Xiaoling Zhang School of Information Science and Engineering Shenyang University of Technology Liaoning Liaohe Laboratory Shenyang Key Laboratory of Advanced Computing and Application Innovation Shenyang China School of Information Science and Engineering Shenyang University of Technology Shenyang China School of Artificial Intelligence Shenyang University of Technology Shenyang Key Laboratory of Industrial Intelligent Chip and Network System Innovation Application Shenyang China

ISBN: (数字)9798331523527

ISBN: (纸本)9798331523534

Artificial intelligence (AI) technology plays a key role in driving the development of a green and smart Energy Internet. As one of the application areas of the Energy Internet, lane detection in intelligent transportation systems can provide accurate real-time traffic data, thus supporting intelligent scheduling and resource optimization, reducing congestion, energy consumption, and emissions, and improving overall energy efficiency. Therefore, deep learning-based lane detection methods can further enhance the intelligence level of the Energy Internet and promote its efficient operation. However, current methods often rely on Convolutional Neural Networks (CNNs), which excel at local feature extraction but struggle to capture global information in complex conditions. In varying light, shadows, and obstacles, CNNs struggle to integrate information, leading to inaccuracies, particularly in dense traffic or complex road structures, where diverse local features result in identification errors and hinder lane continuity. This reduces decision-making efficiency and increases energy consumption. Consequently, we present LFKNet, a lane detection framework designed for the fusion of the Lane Focus Mechanism (LFM) and Kolmogorov-Arnold Network (KAN) to enhance accuracy and robustness. It consists of two key components: the Lane Focus Mechanism , which utilizes the geometrical topology of lanes for improved global fea ture extraction in complex scenarios, and the Kolmogorov-Arnold Network, which employs a learnable nonlinear activation function to optimize performance on complex and nonlinear features. The experiments demonstrate that LFKNet achieves 97.00% accuracy with a 97.76% F1-score on the Tusimple dataset. It also achieves a 79.43% F1-score on the CULane dataset. Overall, LFKNet’s performance validates its effectiveness in lane detection, enabling accurate traffic monitoring and analysis while providing reliable data support for the Energy Internet.

关键词： Energy consumption Accuracy Lane detection Decision making Energy Internet System integration Feature extraction Robustness Topology Artificial intelligence

MAST-Pro: Dynamic Mixture-of-Experts for Adaptive Segmentation of Pan-Tumors with Knowledge-Driven Prompts

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Meng, Runqi Song, Sifan Jin, Pengfei Oh, Yujin Teng, Lin Wang, Yulin Sun, Yiqun Chen, Ling Li, Xiang Li, Quanzheng Guo, Ning Shen, Dinggang School of Biomedical Engineering State Key Laboratory of Advanced Medical Materials and Devices ShanghaiTech University China Center of Advanced Medical Computing and Analysis Massachusetts General Hospital Harvard Medical School Harvard University United States Shanghai United Imaging Intelligence Co. Ltd China Shanghai Clinical Research and Trial Center China

Accurate tumor segmentation is crucial for cancer diagnosis and treatment. While foundation models have advanced general-purpose segmentation, existing methods still struggle with: (1) limited incorporation of medical priors, (2) imbalance between generic and tumor-specific features, and (3) high computational costs for clinical adaptation. To address these challenges, we propose MAST-Pro (Mixture-of-experts for Adaptive Segmentation of pan-Tumors with knowledge-driven Prompts), a novel framework that integrates dynamic Mixture-of-Experts (D-MoE) and knowledge-driven prompts for pan-tumor segmentation. Specifically, text and anatomical prompts provide domain-specific priors, guiding tumor representation learning, while D-MoE dynamically selects experts to balance generic and tumor-specific feature learning, improving segmentation accuracy across diverse tumor types. To enhance efficiency, we employ Parameter-Efficient Fine-Tuning (PEFT), optimizing MAST-Pro with significantly reduced computational overhead. Experiments on multi-anatomical tumor datasets demonstrate that MAST-Pro outperforms state-of-the-art approaches, achieving up to a 5.20% improvement in average DSC while reducing trainable parameters by 91.04%, without compromising accuracy. © 2025, CC BY-NC-ND.

关键词： Image segmentation

Efficient Meta Neural Heuristic for Multi-Objective Combinatorial Optimization

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Chen, Jinbiao Wang, Jiahai Zhang, Zizhen Cao, Zhiguang Ye, Te Chen, Siyuan School of Computer Science and Engineering Sun Yat-sen University China Key Laboratory of Machine Intelligence and Advanced Computing Ministry of Education Sun Yat-sen University China Guangdong Key Laboratory of Big Data Analysis and Processing Guangzhou China School of Computing and Information Systems Singapore Management University Singapore

Recently, neural heuristics based on deep reinforcement learning have exhibited promise in solving multi-objective combinatorial optimization problems (MOCOPs). However, they are still struggling to achieve high learning efficiency and solution quality. To tackle this issue, we propose an efficient meta neural heuristic (EMNH), in which a meta-model is first trained and then fine-tuned with a few steps to solve corresponding single-objective subproblems. Specifically, for the training process, a (partial) architecture-shared multi-task model is leveraged to achieve parallel learning for the meta-model, so as to speed up the training;meanwhile, a scaled symmetric sampling method with respect to the weight vectors is designed to stabilize the training. For the fine-tuning process, an efficient hierarchical method is proposed to systematically tackle all the subproblems. Experimental results on the multi-objective traveling salesman problem (MOTSP), multi-objective capacitated vehicle routing problem (MOCVRP), and multi-objective knapsack problem (MOKP) show that, EMNH is able to outperform the state-of-the-art neural heuristics in terms of solution quality and learning efficiency, and yield competitive solutions to the strong traditional heuristics while consuming much shorter time. Copyright © 2023, The Authors. All rights reserved.

关键词： Reinforcement learning

APRF: Anti-Aliasing Projection Representation Field for Inverse Problem in Imaging

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Chen, Zixuan Yang, Lingxiao Lai, Jianhuang Xie, Xiaohua The School of Computer Science and Engineering Sun Yat-sen University Guangzhou510006 China The Guangdong Province Key Laboratory of Information Security Technology Guangzhou510006 China The Key Laboratory of Machine Intelligence and Advanced Computing Ministry of Education Guangzhou510006 China

Sparse-view Computed Tomography (SVCT) reconstruction is an ill-posed inverse problem in imaging that aims to acquire high-quality CT images based on sparsely-sampled measurements. Recent works use Implicit Neural Representations (INRs) to build the coordinate-based mapping between sinograms and CT images. However, these methods have not considered the correlation between adjacent projection views, resulting in aliasing artifacts on SV sinograms. To address this issue, we propose a self-supervised SVCT reconstruction method – Anti-Aliasing Projection Representation Field (APRF), which can build the continuous representation between adjacent projection views via the spatial constraints. Specifically, APRF only needs SV sinograms for training, which first employs a line-segment sampling module to estimate the distribution of projection views in a local region, and then synthesizes the corresponding sinogram values using a center-based line integral module. After training APRF on a single SV sinogram itself, it can synthesize the corresponding dense-view (DV) sinogram with consistent continuity. High-quality CT images can be obtained by applying re-projection techniques on the predicted DV sinograms. Extensive experiments on CT images demonstrate that APRF outperforms state-of-the-art methods, yielding more accurate details and fewer artifacts. Our code will be publicly available soon. Copyright © 2023, The Authors. All rights reserved.

关键词： Inverse problems

Fine-Grained Depth Knowledge Distillation for Cloth-Changing Person Re-identification

学校读者我要写书评

暂无评论

Fine-Grained Depth Knowledge Distillation for Cloth-Changing...

International Joint Conference on Neural Networks (IJCNN)

作者： Yuhan Yao Jiawei Feng Ancong Wu Jiangqun Ni Wei-Shi Zheng School of Computer Science and Engineering Sun Yat-sen University China School of Cyber Science and Technology Sun Yat-sen University Shenzhen China Department of New Networks Peng Cheng Laboratory Shenzhen China Key Laboratory of Machine Intelligence and Advanced Computing Ministry of Education China Guangdong Key Laboratory of Information Security Technology China

ISBN: (数字)9798350359312

ISBN: (纸本)9798350359329

The mission of cloth-changing person re-identification (CC-ReID) is to discover cloth-invariant and identity-related cues, while traditional person ReID methods rely on appearance features that are biased to cloth-related cues. To tackle this cloth-biased problem, many CC-ReID methods introduced auxiliary body shape information to extract cloth-invariant features, such as 2D sketch images or the 3D Skinned Multi-Person Linear (SMPL) model. However, 2D auxiliary information lacks 3D spatial features, while the 3D SMPL model encounters challenges in capturing features at a finer granularity due to manually defined parameters. To extract fine-grained 3D shape features, we estimate depth maps that contain richer shape information and propose a Fine-grained Depth feature Mining and Distillation (FDMD) framework. We introduce a depth branch and design a fine-grained local feature interaction module to mine fine-grained 3D body shape knowledge from estimated depth maps by exploring the context of semantic-aware local body-part features. To integrate cloth-invariant depth knowledge into the appearance features, the fine-grained 3D shape features are transferred to an appearance branch by feature-space-aligned distillation. Extensive experiments demonstrate that FDMD can achieve state-of-the-art performance on three widely used CC-ReID benchmarks PRCC, Celeb-reID and LaST.

关键词： Solid modeling Three-dimensional displays Shape Neural networks Benchmark testing Feature extraction Data mining

Releasing Inequality Phenomena in L∞-Adversarial Training via Input Gradient Distillation

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Chen, Junxi Dong, Junhao Xie, Xiaohua School of Computer Science and Engineering Sun Yat-sen University Guangzhou510006 China Guangdong Province Key Laboratory of Information Security Technology Guangzhou510006 China Key Laboratory of Machine Intelligence and Advanced Computing Ministry of Education Guangzhou510006 China

Since adversarial examples appeared and showed the catastrophic degradation they brought to DNN, many adversarial defense methods have been devised, among which adversarial training is considered the most effective. However, a recent work showed the inequality phenomena in l∞-adversarial training and revealed that the l∞-adversarially trained model is vulnerable when a few important pixels are perturbed by i.i.d. noise or occluded. In this paper, we propose a simple yet effective method called Input Gradient Distillation (IGD) to release the inequality phenomena in l∞-adversarial training. Experiments show that while preserving the model’s adversarial robustness, compared to PGDAT, IGD decreases the l∞adversarially trained model’s error rate to inductive noise and inductive occlusion by up to 60% and 16.53%, and to noisy images in Imagenet-C by up to 21.11%. Moreover, we formally explain why the equality of the model’s saliency map can improve such robustness. Copyright © 2023, The Authors. All rights reserved.

关键词： Distillation