检索结果-内蒙古大学图书馆

Exploring Token-Level Augmentation in Vision Transformer for Semi-Supervised Semantic Segmentation

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Zhang, Dengke Tang, Quan Liu, Fagui Mei, Haiqing Chen, C.L. Philip The School of Computer Science and Engineering South China University of Technology Guangzhou510006 China The Department of New Network Pengcheng Laboratory Shenzhen518000 China The China Tobacco Guangdong Industrial Corporation Guangzhou510385 China State Key Laboratory of Management and Control for Complex Systems Institute of Automation Chinese Academy of Sciences Beijing100080 China

Semi-supervised semantic segmentation has witnessed remarkable advancements in recent years. However, existing algorithms are based on convolutional neural networks and directly applying them to Vision Transformers poses certain limitations due to conceptual disparities. To this end, we propose TokenMix, a data augmentation technique specifically designed for semi-supervised semantic segmentation with Vision Transformers. TokenMix aligns well with the global attention mechanism by mixing images at the token level, enhancing learning capability for contextual information among image patches. We further incorporate image augmentation and feature augmentation to promote the diversity of augmentation. Moreover, to enhance consistency regularization, we propose a dual-branch framework where each branch applies image and feature augmentation to the input image. We conduct extensive experiments across multiple benchmark datasets, including Pascal VOC 2012, Cityscapes, and COCO. Results suggest that the proposed method outperforms state-of-the-art algorithms with notably observed accuracy improvement, especially under limited fine annotations. Copyright © 2025, The Authors. All rights reserved.

关键词：

Social Force Embedded Mixed Graph Convolutional Network for Multi-class Trajectory Prediction

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Du, Quancheng Wang, Xiao Yin, Shouguo Li, Lingxi Ning, Huansheng The School of Computer and Communication Engineering University of Science and Technology Beijing Beijing100083 China The School of Artificial Intelligence Anhui University Hefei230031 China The State Key Laboratory for Management and Control of Complex Systems Institute of Automation Chinese Academy of Sciences Beijing100190 China The School of Electrical and Computer Engineering Indiana University Purdue University IndianapolisIN46202 United States

Accurate prediction of agent motion trajectories is crucial for autonomous driving, contributing to the reduction of collision risks in human-vehicle interactions and ensuring ample response time for other traffic participants. Current research predominantly focuses on traditional deep learning methods, including convolutional neural networks (CNNs) and recurrent neural networks (RNNs). These methods leverage relative distances to forecast the motion trajectories of a single class of agents. However, in complex traffic scenarios, the motion patterns of various types of traffic participants exhibit inherent randomness and uncertainty. Relying solely on relative distances may not adequately capture the nuanced interaction patterns between different classes of road users. In this paper, we propose a novel multi-class trajectory prediction method named the social force embedded mixed graph convolutional network (SFEM-GCN). The primary goal is to extract social interactions among agents more accurately. SFEM-GCN comprises three graph topologies: the semantic graph (SG), position graph (PG), and velocity graph (VG). These graphs encode various of social force relationships among different classes of agents in complex scenes. Specifically, SG utilizes one-hot encoding of agent-class information to guide the construction of graph adjacency matrices based on semantic information. PG and VG create adjacency matrices to capture motion interaction relationships between different classes agents. These graph structures are then integrated into a mixed graph, where learning is conducted using a spatio-temporal graph convolutional neural network (ST-GCNN). To further enhance prediction performance, we adopt temporal convolutional networks (TCNs) to generate the predicted trajectory with fewer parameters. Experimental results on publicly available datasets demonstrate that SFEM-GCN surpasses state-of-the-art methods in terms of accuracy and robustness. Copyright © 2024, The Authors.

关键词： Forecasting

Robust Visual Question Answering: Datasets, Methods, and Future Challenges

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Ma, Jie Wang, Pinghui Kong, Dechen Wang, Zewei Liu, Jun Pei, Hongbin Zhao, Junzhou Di, Jia The Ministry of Education of Key Laboratory for Intelligent Networks and Network Security School of Cyber Science and Engineering Xi’an Jiaotong University Shaanxi Xi’an710049 China The Ministry of Education of Key Laboratory for Intelligent Networks and Network Security School of Automation Science and Engineering Xi’an Jiaotong University Shaanxi Xi’an710049 China The Shannxi Provincial Key Laboratory of Big Data Knowledge Engineering School of Computer Science and Technology Xi’an Jiaotong University Shaanxi Xi’an710049 China

Visual question answering requires a system to provide an accurate natural language answer given an image and a natural language question. However, it is widely recognized that previous generic VQA methods often tend to memorize biases present in the training data rather than learning proper behaviors, such as grounding images before predicting answers. Therefore, these methods usually achieve high in-distribution but poor out-of-distribution performance. In recent years, various datasets and debiasing methods have been proposed to evaluate and enhance the VQA robustness, respectively. This paper provides the first comprehensive survey focused on this emerging fashion. Specifically, we first provide an overview of the development process of datasets from in-distribution and out-of-distribution perspectives. Then, we examine the evaluation metrics employed by these datasets. Thirdly, we propose a typology that presents the development process, similarities and differences, robustness comparison, and technical features of existing debiasing methods. Furthermore, we analyze and discuss the robustness of representative vision-and-language pre-training models on VQA. Finally, through a thorough review of the available literature and experimental analysis, we discuss the key areas for future research from various viewpoints. Copyright © 2023, The Authors. All rights reserved.

关键词： Visual languages

Stochastic-Robust Planning of Networked Hydrogen-Electrical Microgrids: A Study on Induced Refueling Demand

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Sun, Xunhang Cao, Xiaoyu Zeng, Bo Zhai, Qiaozhu Başar, Tamer Guan, Xiaohong School of Automation Science and Engineering The Ministry of Education Key Laboratory for Intelligent Networks and Network Security Xi’an Jiaotong University Shaanxi Xi’an710049 China Coordinated Science Laboratory University of Illinois Urbana-Champaign UrbanaIL61801 United States Smart Integrated Energy Department Sichuan Digital Economy Industry Development Research Institute Sichuan Chengdu610037 China Department of Industrial Engineering The Department of Electrical and Computer Engineering University of Pittsburgh PittsburghPA15106 United States Center for Intelligent and Networked Systems Department of Automation Tsinghua University Beijing100084 China

Hydrogen-electrical microgrids are increasingly assuming an important role on the pathway toward decarbonization of energy and transportation systems. This paper studies networked hydrogen-electrical microgrids planning (NHEMP), considering a critical but often-overlooked issue, i.e., the demand-inducing effect (DIE) associated with infrastructure development decisions. Specifically, higher refueling capacities will attract more refueling demand of hydrogen-powered vehicles (HVs). To capture such interactions between investment decisions and induced refueling demand, we introduce a decision-dependent uncertainty (DDU) set and build a trilevel stochastic-robust formulation. The upper-level determines optimal investment strategies for hydrogen-electrical microgrids, the lower-level optimizes the risk-aware operation schedules across a series of stochastic scenarios, and, for each scenario, the middle-level identifies the "worst" situation of refueling demand within an individual DDU set to ensure economic feasibility. Then, an adaptive and exact decomposition algorithm, based on Parametric Column-and-Constraint Generation (PC&CG), is customized and developed to address the computational challenge and to quantitatively analyze the impact of DIE. Case studies on an IEEE exemplary system validate the effectiveness of the proposed NHEMP model and the PC&CG algorithm. It is worth highlighting that DIE can make an important contribution to the economic benefits of NHEMP, yet its significance will gradually decrease when the main bottleneck transits to other system restrictions. Copyright © 2024, The Authors. All rights reserved.

关键词： Stochastic systems

SKU-Patch: Towards Efficient Instance Segmentation for Unseen Objects in Auto-Store

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Yang, Biqi Tang, Weiliang Gao, Xiaojie Li, Xianzhi Liu, Yun-Hui Fu, Chi-Wing Heng, Pheng-Ann The Department of Computer Science and Engineering The Chinese University of Hong Kong Hong Kong Guangdong-Hong Kong-Macao Joint Laboratory of Human-Machine Intelligence-Synergy Systems Shenzhen Institutes of Advanced Technology Chinese Academy of Sciences China The School of Computer Science and Technology Huazhong University of Science and Technology China The Department of Mechanical and Automation Engineering The Chinese University of Hong Kong Hong Kong

In large-scale storehouses, precise instance masks are crucial for robotic bin picking but are challenging to obtain. Existing instance segmentation methods typically rely on a tedious process of scene collection, mask annotation, and network fine-tuning for every single Stock Keeping Unit (SKU). This paper presents SKU-Patch, a new patch-guided instance segmentation solution, leveraging only a few image patches for each incoming new SKU to predict accurate and robust masks, without tedious manual effort and model re-training. Technical-wise, we design a novel transformer-based network with (i) a patch-image correlation encoder to capture multi-level image features calibrated by patch information and (ii) a patch-aware transformer decoder with parallel task heads to generate instance masks. Extensive experiments on four storehouse benchmarks manifest that SKU-Patch is able to achieve the best performance over the state-of-the-art methods. Also, SKU-Patch yields an average of nearly 100% grasping success rate on more than 50 unseen SKUs in a robot-aided auto-store logistic pipeline, showing its effectiveness and practicality. Copyright © 2023, The Authors. All rights reserved.

关键词： Benchmarking

TopoSeg: Topology-Aware Nuclear Instance Segmentation

学校读者我要写书评

暂无评论

TopoSeg: Topology-Aware Nuclear Instance Segmentation

International Conference on computer Vision (ICCV)

作者： Hongliang He Jun Wang Pengxu Wei Fan Xu Xiangyang Ji Chang Liu Jie Chen School of Electronic and Computer Engineering Peking University Shenzhen China School of Computer Science and Technology Soochow University Suzhou China Sun Yat-sen University Guangzhou China Peng Cheng Laboratory Shenzhen China Department of Automation and BNRist Tsinghua University Beijing China

Nuclear instance segmentation has been critical for pathology image analysis in medical science, e.g., cancer diagnosis. Current methods typically adopt pixel-wise optimization for nuclei boundary exploration, where rich structural information could be lost for subsequent quantitative morphology assessment. To address this issue, we develop a topology-aware segmentation approach, termed TopoSeg, which exploits topological structure information to keep the predictions rational, especially in common situations with densely touching and overlapping nucleus instances. Concretely, TopoSeg builds on a topology-aware module (TAM), which encodes dynamic changes of different topology structures within the three-class probability maps (inside, boundary, and background) of the nuclei to persistence barcodes and makes the topology-aware loss function. To efficiently focus on regions with high topological errors, we propose an adaptive topology-aware selection (ATS) strategy to enhance the topology-aware optimization procedure further. Experiments on three nuclear instance segmentation datasets justify the superiority of TopoSeg, which achieves state-of-the-art performance. The code is available at https://***/hhlisme/toposeg.

关键词：

Robust Iterative Learning Control Design for a Class of Uncertain Batch Processes

学校读者我要写书评

暂无评论

Robust Iterative Learning Control Design for a Class of Unce...

European Control Conference (ECC)

作者： Robert Maniarski Wojciech Paszke Hongfeng Tao Eric Rogers Institute of Automation Electronic and Electrical Engineering University of Zielona Góra Zielona Góra Poland Key Laboratory of Advanced Process Control for Light Industry of Ministry of Education Jiangnan University Wuxi China School of Electronics and Computer Science University of Southampton Southampton UK

ISBN: (数字)9783907144107

ISBN: (纸本)9798331540920

This paper investigates the problem of designing robust iterative learning control laws for discrete-time batch processes with norm-bounded parameter uncertainties. A law of proportional-differential type is designed to achieve robust convergence of the tracking error in the batch-to-batch direction. It is shown that the design problem can be written as a two-dimensional system. Then, the recently developed non-conservative conditions for (structural) stability analysis for a linear Roesser model are used. The conditions for the existence and computation of the required control law matrices are linear matrix inequality-based. Finally, comparative simulation results show the effectiveness of the new design.

关键词： Uncertain systems Analytical models Simulation Computational modeling Batch production systems Europe Linear matrix inequalities

Neural Dialogue Model with Retrieval Attention for Personalized Response Generation

学校读者我要写书评

暂无评论

computers, Materials & Continua 2020年第1期62卷 113-122页

作者： Cong Xu Zhenqi Sun Qi Jia Dezheng Zhang Yonghong Xie Alan Yang School of Automation and Electrical Engineering University of Science and Technology BeijingBeijing100083China Beijing Key Laboratory of Knowledge Engineering for Materials Science University of Science and Technology BeijingBeijing100083China School of Computer and Communication Engineering University of Science and Technology BeijingBeijing100083China Amphenol AssembleTech HoustonTX 77070US

With the success of new speech-based human-computer interfaces,there is a great need for effective and friendly dialogue agents that can communicate with people naturally and ***,the lack of personality and consistency is one of critical problems in neural dialogue *** this paper,we aim to generate consistent response with fixed profile and background information for building a realistic dialogue *** on the encoder-decoder model,we propose a retrieval mechanism to deliver natural and fluent response with proper information from a profile ***,in order to improve the efficiency of training the dataset related to profile information,we adopt a method of pre-training and adjustment for general dataset and profile *** model is trained by social dialogue data from *** to both automatic and human evaluation metrics,the proposed model significantly outperforms standard encoder-decoder model and other improved models on providing the correct profile and high-quality responses.

关键词： Dialogue system LSTM encoder-decoder model attention mechanism

UGNA-VPR: A Novel Training Paradigm for Visual Place Recognition Based on Uncertainty-Guided NeRF Augmentation

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Shen, Yehui Zhang, Lei Li, Qingqiu Zhao, Xiongwei Wang, Yue Lu, Huimin Chen, Xieyuanli College of Intelligence Science and Technology National University of Defense Technology China Faculty of Robot Science and Engineering Northeastern University China School of Electronic and Information Engineering Harbin Institute of Technology Shenzhen China State Key Laboratory of Industrial Control Technology Institute of Cyber-Systems and Control Zhejiang University China SIASUN Robot& Automation Co. Ltd China School of Computer Science Fudan University China

Visual place recognition (VPR) is crucial for robots to identify previously visited locations, playing an important role in autonomous navigation in both indoor and outdoor environments. However, most existing VPR datasets are limited to single-viewpoint scenarios, leading to reduced recognition accuracy, particularly in multi-directional driving or feature-sparse scenes. Moreover, obtaining additional data to mitigate these limitations is often expensive. This paper introduces a novel training paradigm to improve the performance of existing VPR networks by enhancing multi-view diversity within current datasets through uncertainty estimation and NeRF-based data augmentation. Specifically, we initially train NeRF using the existing VPR dataset. Then, our devised self-supervised uncertainty estimation network identifies places with high uncertainty. The poses of these uncertain places are input into NeRF to generate new synthetic observations for further training of VPR networks. Additionally, we propose an improved storage method for efficient organization of augmented and original training data. We conducted extensive experiments on three datasets and tested three different VPR backbone networks. The results demonstrate that our proposed training paradigm significantly improves VPR performance by fully utilizing existing data, outperforming other training approaches. We further validated the effectiveness of our approach on self-recorded indoor and outdoor datasets, consistently demonstrating superior results. Our dataset and code have been released at https://***/nubot-nudt/UGNA-VPR. Copyright © 2025, The Authors. All rights reserved.

关键词：