检索结果-内蒙古大学图书馆

Mining Unseen Classes via Regional Objectness: A Simple Baseline for Incremental Segmentation

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Zhang, Zekang Gao, Guangyu Fang, Zhiyuan Jiao, Jianbo Wei, Yunchao School of Computer Science Beijing Institute of Technology China School of Computer Science University of Birmingham United Kingdom WEI Lab Institute of Information Science Beijing Jiaotong University China Beijing Key Laboratory of Advanced Information Science and Network China

Incremental or continual learning has been extensively studied for image classification tasks to alleviate catastrophic forgetting, a phenomenon that earlier learned knowledge is forgotten when learning new concepts. For class incremental semantic segmentation, such a phenomenon often becomes much worse due to the background shift, i.e., some concepts learned at previous stages are assigned to the background class at the current training stage, therefore, significantly reducing the performance of these old concepts. To address this issue, we propose a simple yet effective method in this paper, named Mining unseen Classes via Regional Objectness for Segmentation (MicroSeg). Our MicroSeg is based on the assumption that background regions with strong objectness possibly belong to those concepts in the historical or future stages. Therefore, to avoid forgetting old knowledge at the current training stage, our MicroSeg first splits the given image into hundreds of segment proposals with a proposal generator. Those segment proposals with strong objectness from the background are then clustered and assigned newly-defined labels during the optimization. In this way, the distribution characterizes of old concepts in the feature space could be better perceived, relieving the catastrophic forgetting caused by the background shift accordingly. Extensive experiments on Pascal VOC and ADE20K datasets show competitive results with state-of-the-art, well validating the effectiveness of the proposed MicroSeg. Code is available at https://***/zkzhang98/MicroSeg. © 2022, CC BY.

关键词： Semantics

Tile Classification Based Viewport Prediction with Multi-modal Fusion Transformer

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Zhang, Zhihao Chen, Yiwei Zhang, Weizhan Yan, Caixia Zheng, Qinghua Wang, Qi Chen, Wangdu School of Computer Science and Technology Xi'an Jiaotong University Key Laboratory of Intelligent Networks and Network Security Ministry of Education Xi'an Jiaotong University Xi’an China School of Computer Science and Technology Xi'an Jiaotong University & National Engineering Lab for Big Data Analytics Xi'an Jiaotong University Xi’an China MIGU Video Shanghai China

Viewport prediction is a crucial aspect of tile-based 360◦ video streaming system. However, existing trajectory based methods lack of robustness, also oversimplify the process of information construction and fusion between different modality inputs, leading to the error accumulation problem. In this paper, we propose a tile classification based viewport prediction method with Multi-modal Fusion Transformer, namely MFTR. Specifically, MFTR utilizes transformer-based networks to extract the long-range dependencies within each modality, then mine intra- and inter-modality relations to capture the combined impact of user historical inputs and video contents on future viewport selection. In addition, MFTR categorizes future tiles into two categories: user interested or not, and selects future viewport as the region that contains most user interested tiles. Comparing with predicting head trajectories, choosing future viewport based on tile's binary classification results exhibits better robustness and interpretability. To evaluate our proposed MFTR, we conduct extensive experiments on two widely used PVS-HM and Xu-Gaze dataset. MFTR shows superior performance over state-of-the-art methods in terms of average prediction accuracy and overlap ratio, also presents competitive computation efficiency. © 2023, CC BY.

关键词： Forecasting

A Method of Shadowsocks(R) Traffic Identification Based on Protocol Analysis 21

学校读者我要写书评

暂无评论

A Method of Shadowsocks(R) Traffic Identification Based on P...

21st IEEE International Conference on Communication technology, ICCT 2021

作者： Luo, Jie Bao, Liang Ni, Lvlin The 30th Research Institute of China Electronics Technology Group Corporation Science and Technology on Communication Security Laboratory Sichuan Chengdu China The Third Research Institute of Ministry of Public Security Key Lab of Information Network Security of Ministry of Public Security Shanghai China

ISBN: (纸本)9781665432061

Shadowsocks (R) is a proxy software based on Socks5, which is the collective name of Shadowsocks and Shadowsocks (R). In order to improves the shortcomings of other softwares breaking internet censorship, Shadowsocks (R) is a private protocol without handshake negotiation mechanism, and pays more attention to high camouflage and high concealment of traffic, which makes it difficult to use common methods(fingerprint identification, port filtering)to identify Shadowsocks (R) traffic[1], This paper proposes a method of Shadowsocks (R) traffic identification from the perspective of the protocol, which extracts features from four dimensions, and inputs features into the machine learning model for training. This method performs secondary identification and statistical filtering on the output of the machine learning model, and finally identifys Shadowsocks (R) traffic. This method has been very well validated in the laboratory environment. © 2021 IEEE.

关键词： Machine learning

Hard Negatives or False Negatives: Correcting Pooling Bias in Training Neural Ranking Models

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Cai, Yinqiong Guo, Jiafeng Fan, Yixing Ai, Qingyao Zhang, Ruqing Cheng, Xueqi CAS Key Lab of Network Data Science and Technology ICT CAS University of Chinese Academy of Sciences Beijing China Dept. CS&T Beijing National Research Center for Information Science and Technology Tsinghua University Beijing China

Neural ranking models (NRMs) have become one of the most important techniques in information retrieval (IR). Due to the limitation of relevance labels, the training of NRMs heavily relies on negative sampling over unlabeled data. In general machine learning scenarios, it has shown that training with hard negatives (i.e., samples that are close to positives) could lead to better performance. Surprisingly, we find opposite results from our empirical studies in IR. When sampling top-ranked results (excluding the labeled positives) as negatives from a stronger retriever, the performance of the learned NRM becomes even worse. Based on our investigation, the superficial reason is that there are more false negatives (i.e., unlabeled positives) in the top-ranked results with a stronger retriever, which may hurt the training process;The root is the existence of pooling bias in the dataset constructing process, where annotators only judge and label very few samples selected by some basic retrievers. Therefore, in principle, we can formulate the false negative issue in training NRMs as learning from labeled datasets with pooling bias. To solve this problem, we propose a novel Coupled Estimation Technique (CET) that learns both a relevance model and a selection model simultaneously to correct the pooling bias for training NRMs. Empirical results on three retrieval benchmarks show that NRMs trained with our technique can achieve significant gains on ranking effectiveness against other baseline strategies. © 2022, CC0.

关键词：

Modulating p-type doping of two-dimensional material palladium diselenide

学校读者我要写书评

暂无评论

Nano Research 2024年第4期17卷 3232-3244页

作者： Jiali Yang Yu Liu En-Yang Wang Jinbo Pang Shirong Huang Thomas Gemming Jinshun Bi Alicja Bachmatiuk Hao Jia Shu-Xian Hu Chongyun Jiang Hong Liu Gianaurelio Cuniberti Weijia Zhou Mark H Rümmeli Institute for Advanced Interdisciplinary Research(iAIR) University of JinanJinan 250022China Institute for Materials Chemistry Leibniz Institute for Solid State and Materials Research Dresden(IFW Dresden)20 Helmholtz StrasseDresden 01069Germany State Key Lab of Transducer Technology Shanghai Institute of Microsystem and Information TechnologyChinese Academy of SciencesShanghai 200050China College of Energy Soochow Institute for Energy and Materials Innovations Soochow UniversitySuzhou 215006China Key Laboratory of Advanced Carbon Materials and Wearable Energy Technologies of Jiangsu Province Soochow UniversitySuzhou 215006China School of Integrated Circuits University of Chinese Academy of SciencesBeijing 101408China Institute of Microelectronics of Tianjin Binhai New Area Tianjin 300451China Institute of Microelectronics Chinese Academy of SciencesBeijing 100029China Lukasiewicz Research Network PORT Polish Center for Technology DevelopmentStablowicka 147Wroclaw 54-066Poland School of Mathematics and Physics University of Science and Technology BeijingBeijing 100083China College of electronic information and optical engineering Nankai UniversityTianjin 300350China Centre of Polymer and Carbon Materials Polish Academy of SciencesM.Curie Sklodowskiej 34Zabrze 41-819Poland Center for Energy and Environmental Technologies VŠB-Technical University of Ostrava17.Listopadu 15Ostrava 70833Czech Republic State Key Laboratory of Crystal Materials Center of Bio&Micro/Nano Functional MaterialsShandong UniversityJinan 250100China Dresden Center for Computational Materials Science Technische Universität DresdenDresden 01062Germany Dresden Center for Intelligent Materials(GCL DCIM) Technische Universität DresdenDresden 01062Germany Institute for Materials Science and Max Bergmann Center of Biomaterials Technische Universität DresdenDresden 01069Germany Center for Advancing Electronics Dresden Technische Universität DresdenDresden 01069Germany

The van der Waals heterostructures have evolved as novel materials for complementing the Si-based semiconductor ***-10 noble metal dichalcogenides(e.g.,PtS_(2),PtSe_(2),PdS_(2),and PdSe_(2))have been listed into two-dimensional(2D)materials toolkit to assemble van der Waals *** them,PdSe_(2) demonstrates advantages of high stability in air,high mobility,and wide tunable ***,the regulation of p-type doping of PdSe_(2) remains unsolved problem prior to fabricating p–n junction as a fundamental platform of semiconductor ***,a quantitative method for the controllable doping of PdSe_(2) is yet to be *** this study,the doping level of PdSe_(2) was correlated with the concentration of Lewis acids,for example,SnCl_(4),used for *** the transfer characteristics,the threshold voltage(the gate voltage corresponding to the minimum drain current)increased after SnCl_(4) soaking ***_(2) transistors were soaked in SnCl_(4) solutions with five different *** threshold voltages from the as-obtained transfer curves were extracted for linear fitting to the threshold voltage versus doping concentration correlation *** study provides in-depth insights into the controllable p-type doping of PdSe_(2).It may also push forward the research of the regulation of conductivity behaviors of 2D materials.

关键词： two-dimensional(2D)materials Lewis acid treatment p-type doping field-effect transistors transfer characteristic

Mamba-Sea: A Mamba-based Framework with Global-to-Local Sequence Augmentation for Generalizable Medical Image Segmentation

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Cheng, Zihan Guo, Jintao Zhang, Jian Qi, Lei Zhou, Luping Shi, Yinghuan Gao, Yang Shanghai Jiao Tong University School of Medicine Shanghai200025 China National Institute of Healthcare Data Science Nanjing University Jiangsu Nanjing210093 China State Key Laboratory for Novel Software Technology the National Institute of Healthcare Data Science Nanjing University Jiangsu Nanjing210093 China School of Computer Science and Engineering Key Lab of Computer Network and Information Integration Southeast University Jiangsu Nanjing211189 China School of Electrical and Information Engineering The University of Sydney SydneyNSW2006 Australia

To segment medical images with distribution shifts, domain generalization (DG) has emerged as a promising setting to train models on source domains that can generalize to unseen target domains. Existing DG methods are mainly based on CNN or ViT architectures. Recently, advanced state space models, represented by Mamba, have shown promising results in various supervised medical image segmentation. The success of Mamba is primarily owing to its ability to capture long-range dependencies while keeping linear complexity with input sequence length, making it a promising alternative to CNNs and ViTs. Inspired by the success, in the paper, we explore the potential of the Mamba architecture to address distribution shifts in DG for medical image segmentation. Specifically, we propose a novel Mamba-based framework, Mamba-Sea, incorporating global-to-local sequence augmentation to improve the model’s generalizability under domain shift issues. Our Mamba-Sea introduces a global augmentation mechanism designed to simulate potential variations in appearance across different sites, aiming to suppress the model’s learning of domain-specific information. At the local level, we propose a sequence-wise augmentation along input sequences, which perturbs the style of tokens within random continuous sub-sequences by modeling and resampling style statistics associated with domain shifts. To our best knowledge, Mamba-Sea is the first work to explore the generalization of Mamba for medical image segmentation, providing an advanced and promising Mamba-based architecture with strong robustness to domain shifts. Remarkably, our proposed method is the first to surpass a Dice coefficient of 90% on the Prostate dataset, which exceeds previous SOTA of 88.61%. The code is available at https://***/orange-czh/Mamba-Sea. Copyright © 2025, The Authors. All rights reserved.

关键词： Image segmentation

Multiple asymmetric couplings induced unconventional corner mode in topolectrical circuits

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Jiang, Hengxuan Wang, Xiumei Chen, Jie Zhou, Xingping College of Integrated Circuit Science and Engineering Nanjing University of Posts and Telecommunications Nanjing210003 China College of Electronic and Optical Engineering Nanjing University of Posts and Telecommunications Nanjing210003 China Institute of Quantum Information and Technology Key Lab of Broadband Wireless Communication and Sensor Network Technology Ministry of Education Nanjing University of Posts and Telecommunications Nanjing210003 China

We investigate the emergence of unconventional corner mode in a two-dimensional topolectrical circuits induced by asymmetric couplings. The non-Hermitian skin effect of two kinked one-dimensional lattices with multiple asymmetric couplings are explored. Then we extend to the two-dimensional model, derive conditions for the non-Hermitian hybrid skin effect and show how the corner modes are formed by nonreciprocal pumping based on one-dimensional topological modes. We provide explicit electrical circuit setups for realizing our observations via realistic LTspice simulation. Moreover, we show the time varying behaviors of voltage distributions to confirm our results. Our study may help to extend the knowledge on building the topological corner modes in the non-Hermitian presence. © 2024, CC BY.

关键词： Topology

Finformer: A Static-dynamic Spatiotemporal Framework for Stock Trend Prediction

学校读者我要写书评

暂无评论

Finformer: A Static-dynamic Spatiotemporal Framework for Sto...

IEEE International Conference on Big Data

作者： Yi Zu Jiacong Mi Lingning Song Shan Lu Jieyue He School of Computer Science and Engineering Key Lab of Computer Network and Information Integration MOE Southeast University Nanjing China School of Software Engineering Southeast University Nanjing China Nanjing Fenghuo Tiandi Communication Technology Co. Ltd Nanjing China

The core of quantitative investment lies in predicting future trends in stock prices. The future trend of a stock is closely related to the industry it belongs to and its relationship with other stocks. Although some research has focused on stock trend prediction in recent years, most studies have only considered the stock’s own time series feature, neglecting the spatial features between stocks. Some research has incorporated spatial information, but typically only considered predefined static relationships. At the same time, capturing dynamic spatial information in the market has been a long-standing challenge. Thus, we propose a spatio-temporal model, Finformer, in order to go beyond traditional time series models. We designed a sparse static-dynamic transformer to capture dynamic market spatial information as it changes over time and combined predefined relationships to extract highly correlated spatial features in the stock market. To effectively integrate spatial and temporal features, we introduced an adaptive spatio-temporal fusion module that dynamically fuses spatio-temporal features based on market conditions at different periods. Experiments on two real-world stock market datasets show that our proposed model outperforms the state-of-the-art baselines in the signal-based and portfolio-based metrics, which are widely concerned in the financial field. Ablation study and hyper-parameter study further reveal the effectiveness of each module in the model and the impact of hyper-parameters. The code will be made publicly available. 1

关键词：

Modeling topical relevance for multi-turn dialogue generation 29

学校读者我要写书评

暂无评论

Modeling topical relevance for multi-turn dialogue generatio...

29th International Joint Conference on Artificial Intelligence, IJCAI 2020

作者： Zhang, Hainan Lan, Yanyan Pang, Liang Chen, Hongshen Ding, Zhuoye Yin, Dawei CAS Key Lab of Network Data Science and Technology Institute of Computing Technology CAS China *** Beijing China University of Chinese Academy of Sciences China

ISBN: (纸本)9780999241165

Topic drift is a common phenomenon in multi-turn dialogue. Therefore, an ideal dialogue generation models should be able to capture the topic information of each context, detect the relevant context, and produce appropriate responses accordingly. However, existing models usually use word or sentence level similarities to detect the relevant contexts, which fail to well capture the topical level relevance. In this paper, we propose a new model, named STAR-BTM, to tackle this problem. Firstly, the Biterm Topic Model is pre-trained on the whole training dataset. Then, the topic level attention weights are computed based on the topic representation of each context. Finally, the attention weights and the topic distribution are utilized in the decoding process to generate the corresponding responses. Experimental results on both Chinese customer services data and English Ubuntu dialogue data show that STAR-BTM significantly outperforms several state-of-the-art methods, in terms of both metric-based and human evaluations. © 2020 Inst. Sci. inf., Univ. Defence in Belgrade. All rights reserved.

关键词： Stars