检索结果-内蒙古大学图书馆

SSRN 2022年

作者： Liu, Haoliang Zhang, Yu School of Computer Science and Engineering The Key Lab of Computer Network and Information Integration Ministry of Education Southeast University Nanjing211189 China

In this paper, we study the problem of image-text matching. Image-text matching aims to train an model that learns to identify whether or not an (image, caption) input pair constitutes an aligned pair. To boost image-text matching, the soft label is introduced by knowledge distillation. Its benefit is generally attributed to the privileged information on similarities among image-text pairs provided by the teacher model. However, training a stronger teacher is computationally heavy and inefficient. In this paper, we propose a novel parameter-free Soft Labels Generation(SLG) method for image-text matching which can produce soft labels with privileged information without the need of training a heavy teacher model. We propose to use intra-modal knowledge to generate the soft label for each image-text pair, which provides an alternative to knowledge distillation. Our method is especially applicable to cases where a stronger teacher model is not available, or only limited computation resources are provided. Systematic experiments conducted on public benchmarks demonstrate that our SLG is empirically effective. Our SLG achieves about 2%-3% recall@1 boosting than the baseline. © 2022, The Authors. All rights reserved.

关键词： Personnel training

来源：评论

学校读者我要写书评

暂无评论

DLFaaS:Serverless Platform for Data-Intensive Tasks Based on Interval Access Patterns

DLFaaS:Serverless Platform for Data-Intensive Tasks Based on...

引用

IEEE/ACM Int'l Conference on & Int'l Conference on Cyber, Physical and Social Computing (CPSCom) Green Computing and Communications (GreenCom)

作者： Yang Cao Wenbin Song Hanqian Wu Shengchao Yuan School of Computer Science and Engineering Southeast University China School of Cyber Science and Engineering Southeast University China Key Laboratory of Computer Network and Information Integration of Ministry of Education Southeast University China

Serverless architecture is a novel paradigm in cloud computing. The increasing trend in applications is shifting towards serverless platforms to achieve on-demand payment and elastic scalability. However, for data-intensive functions, using serverless architectures can result in performance losses due to frequent data transfers, incurring significant time overhead for data movement. Current serverless systems lack optimization for data migration times in tasks involving data-intensive functions. We conducted an analysis of serverless architecture, function invocation patterns, and latency for data-intensive functions. We found that the lack of awareness of access patterns for these functions leads to inefficient cache eviction, resulting in low cache hit rates. Additionally, bursty calls result in repeated remote storage accesses, contributing to time overhead in dataintensive ***, we proposed a serverless platform based on function invocation patterns, optimized for data-intensive tasks. Specifically, we designed a two-tier caching queue, where data objects accessed multiple times for read and write operations were cached based on historical access patterns. These objects were placed in different cache lists according to various time intervals. We employed an eviction strategy based on function access intervals to enhance overall hit rates. Compared to popular existing solutions, we reduced time latency by 32.1% and improved cache hit rates by 12%.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Image-Text Retrieval with Intra-Modal Soft Labels

SSRN

引用

SSRN 2022年

作者： Liu, Haoliang Zhang, Yu School of Computer Science and Engineering The Key Lab of Computer Network and Information Integration Ministry of Education Southeast University Nanjing211189 China

Image-text retrieval seeks to find relevant images from a huge image corpus given a textual description, and vice versa. The critical challenge of this task is to well learn image-text alignments to measure the image-text similarity. Existing works normally label the negative/positive image-text pair with a binary label. They assume that a text is only relevant to the paired images and irrelevant to other images in the corpus. However, using the binary labels during training may be sub-optimal for several reasons. First, a binary label can not reveal the precise similarity between an image and a text. Second, the annotated labels may contain noise. In this paper, we study the idea of training the model with the soft label information in which each image-text pair is associated with a specific soft label. Intuitively, if the textual description of an image shares a high textual similarity with a text, the image should also be highly relevant to the text. Therefore, we determine the soft label between the cross-modal (image, text) pair by the similarity between the intra-modal (image caption, text) pair, which is termed as intra-modal soft label (ISL). In the experiment, we evaluated our method on benchmark datasets, which demonstrate the effectiveness of the proposed method. © 2022, The Authors. All rights reserved.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Topology Uncertainty Modeling For Imbalanced Node Classification on Graphs

Topology Uncertainty Modeling For Imbalanced Node Classifica...

引用

International Conference on Acoustics, Speech, and Signal Processing (ICASSP)

作者： Jiayi Gao Jiaxing Li Ke Zhang Youyong Kong Jiangsu Provincial Joint International Research Laboratory of Medical Information Processing School of Computer Science and Engineering Southeast University Nanjing China Ministry of Education Key Laboratory of Computer Network and Information Integration Southeast University Nanjing China

Most existing graph neural networks work under a class-balanced assumption, while ignoring class-imbalanced scenarios that widely exist in real-world graphs. Although there are many methods in other fields that can alleviate this issue, they do not consider the special topology of the non-Euclidean graph. Hence, we propose Graph Topology Uncertainty (GraphTU), a novel probabilistic class-imbalanced solution specifically for graphs. Firstly, an invisible "uncertain gap" between under-represented minorities in training set and authentic minorities in unseen set is modeled by estimating statistical variances in topology. We extend the training distribution for minorities by sampling in this gap through a non-parametric way. Moreover, a gradient-guided mask is introduced to prevent biased statistics. Extensive experiments demonstrate the superior performance of GraphTU.

关键词： Training Uncertainty network topology Signal processing Probabilistic logic Graph neural networks Acoustics

来源：评论

学校读者我要写书评

暂无评论

Topgformer: Topological-Based Graph Transformer for Mapping Brain Structural Connectivity to Functional Connectivity

Topgformer: Topological-Based Graph Transformer for Mapping ...

引用

International Conference on Acoustics, Speech, and Signal Processing (ICASSP)

作者： Dalu Guo Ke Zhang Jiaxing Li Youyong Kong Jiangsu Provincial Joint International Research Laboratory of Medical Information Processing School of Computer Science and Engineering Southeast University Nanjing China Key Laboratory of Computer Network and Information Integration Southeast University Ministry of Education Nanjing China

Exploring the mapping between structural connectivity (SC) and functional connectivity (FC) is of essential importance to understanding the working mechanism of the human brain. Traditional methods are difficult to represent the complex relationship of high-order interaction between SC and FC. Recent learning-based methods can not well capture the important long-range interactions and edge information of brain connectivity. To address the issue, we propose a novel Topological-based Graph Transformer (TopGFormer) to generate functional connectivity from the structure connectivity with sufficient consideration of topological properties of brain connectivity. We propose a Topological Multi-Head Attention block that simultaneously uses centrality encoding and adjacency matrix to capture node and edge importance respectively. The centrality encoding reflects the importance of brain regions while the adjacency matrix captures the linking relationship of different regions. Experiments on HCP resting and emotion task datasets demonstrate the performance of the proposed method.

关键词： Learning systems Brain Correlation Signal processing Transformers Encoding Acoustics

来源：评论

学校读者我要写书评

暂无评论

RoKEPG: RoBERTa and Knowledge Enhancement for Prescription Generation of Traditional Chinese Medicine

RoKEPG: RoBERTa and Knowledge Enhancement for Prescription G...

引用

IEEE International Conference on Bioinformatics and Biomedicine (BIBM)

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

Traditional Chinese medicine (TCM) prescription is the most critical form of TCM treatment, and uncovering the complex nonlinear relationship between symptoms and TCM is of great significance for clinical practice and assisting physicians in diagnosis and treatment. Although there have been some studies on TCM prescription generation, these studies consider a single factor and directly model the symptom-prescription generation problem mainly based on symptom descriptions, lacking guidance from TCM knowledge. To this end, we propose a RoBERTa and Knowledge Enhancement model for Prescription Generation of Traditional Chinese Medicine (RoKEPG). RoKEPG is firstly pre-trained by our constructed TCM corpus, followed by fine-tuning the pre-trained model, and the model is guided to generate TCM prescriptions by introducing four classes of knowledge of TCM through the attention mask matrix. Experimental results on the publicly available TCM prescription dataset show that RoKEPG improves the F 1 metric by about 2% over the baseline model with the best results.

关键词：

来源：评论

学校读者我要写书评

暂无评论

SAM-Event-Adapter: Adapting Segment Anything Model for Event-RGB Semantic Segmentation

SAM-Event-Adapter: Adapting Segment Anything Model for Event...

引用

IEEE International Conference on Robotics and Automation (ICRA)

作者： Bowen Yao Yongjian Deng Yuhan Liu Hao Chen Youfu Li Zhen Yang College of Computer Science Beijing University of Technology Beijing China Key Lab of Computer Network and Information Integration (Southeast University) Ministry of Education Nanjing China Department of Mechanical Engineering City University of Hong Kong Kowloon Hong Kong SAR

ISBN: (数字)9798350384574

ISBN: (纸本)9798350384581

Semantic segmentation, a fundamental visual task ubiquitously employed in sectors ranging from transportation and robotics to healthcare, has always captivated the research community. In the wake of rapid advancements in large model research, the foundation model for semantic segmentation tasks, termed the Segment Anything Model (SAM), has been introduced. This model substantially addresses the dilemma of poor generalizability of previous segmentation models and the disadvantage in requiring to retrain the whole model on variant datasets. Nonetheless, segmentation models developed on SAM remain constrained by the inherent limitations of RGB sensors, particularly in scenarios characterized by complex lighting conditions and high-speed motion. Motivated by these observations, a natural recourse is to adapt SAM to additional visual modalities without compromising its robust generalizability. To achieve this, we introduce a lightweight SAM-Event-Adapter (SE-Adapter) module, which incorporates event camera data into a cross-modal learning architecture based on SAM, with only limited tunable parameters incremental. Capitalizing on the high dynamic range and temporal resolution afforded by event cameras, our proposed multi-modal Event-RGB learning architecture effectively augments the performance of semantic segmentation tasks. In addition, we propose a novel paradigm for representing event data in a patch format compatible with transformer-based models, employing multi-spatiotemporal scale encoding to efficiently extract motion and semantic correlations from event representations. Exhaustive empirical evaluations conducted on the DSEC-Semantic and DDD17 datasets provide validation of the effectiveness and rationality of our proposed approach.

关键词： Adaptation models Visualization Semantic segmentation Motion segmentation Transportation Cameras Transformers

来源：评论

学校读者我要写书评

暂无评论

Timo: In-memory temporal query processing for big temporal data

Timo: In-memory temporal query processing for big temporal d...

引用

作者： Zheng, Xiao Liu, Houkai Wang, Xiujun Wu, Xuangou Yu, Feng School of Computer Science and Technology Anhui University of Technology Maanshan China Institute for Artificial Intelligence Hefei Comprehensive National Science Center Hefei China Anhui Engineering Laboratory for Intelligent Applications and Security of Industrial Internet Anhui University of Technology Anhui Maanshan China School of Computer Science and Engineering & School of Software Guangxi Normal University Guilin China Key Laboratory of Computer Network and Information Integration Southeast University Nanjing China

Today's internet applications generate massive temporal data anywhere and anytime. Although some disk-based temporal systems are currently available, they suffer from poor I/O performance, especially when applied in intelligent applications deployed in the cloud and edge environments. Therefore, how to process temporal operations with low latency and high throughput becomes a crucial problem for efficient data processing. This paper proposes Timo, a distributed in-memory temporal query and analytic model for big temporal data. Firstly, a space-efficient temporal index is proposed to support more efficiently query performance with less memory space than the state-of-art methods. Secondly, based on the temporal locality feature of temporal queries, Timo proposes a partitioner mechanism, which utilizes forward scan algorithm, to improve query throughput. Thirdly, some optimal strategies are proposed to improve the execution process of temporal query and analysis, which reduce intermediate result data size and improve the throughput much further. Lastly, we implement the Timo system on the Apache Spark platform, which extends Spark dataset API with temporal query and analysis API functions for users. Extensive experimental results show that the Timo system outperforms other Spark-based temporal systems in terms of both query latency and throughput. © 2021 John Wiley & Sons, Ltd.

关键词： Data handling

来源：评论

学校读者我要写书评

暂无评论

Adaptive group personalization for federated mutual transfer learning 24

Adaptive group personalization for federated mutual transfer...

引用

Proceedings of the 41st International Conference on Machine Learning

作者： Haoqing Xu Dian Shen Meng Wang Beilun Wang School of Computer Science and Engineering Southeast University Nanjing China School of Computer Science and Engineering Southeast University Nanjing China and Key Laboratory of Computer Network and Information Integration (Southeast University) Ministry of Education China College of Design and Innovation Tongji University Shanghai China

Mutual transfer learning aims to improve prediction with knowledge from related domains. Recently, federated learning is applied in this field to address the communication and privacy concerns. However, previous clustered federated learning (CFL) solutions lack theoretical guarantee of learnability recovery and require time-consuming hyper-parameter tuning, while centralized mutual transfer learning methods lack adaptability to concept drifts. In this paper, we propose the Adaptive Group Personalization method (AdaGrP) to overcome these challenges. We adaptively decide the recovery threshold with a non-parametric method, adaptive threshold correction, for tuning-free solution with relaxed condition. Theoretical results guarantee the perfect learnability recovery with the corrected threshold. Empirical results show AdaGrP achieves 16.9% average improvement in learnability structure recovery compared with state-of-the-art CFL baselines.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Transformer-based Multi-Instance Learning for Weakly Supervised Object Detection

arXiv

引用

arXiv 2023年

作者： Wang, Zhaofei Zhang, Weijia Zhang, Min-Ling School of Computer Science and Engineering Southeast University Nanjing210096 China Key Laboratory of Computer Network and Information Integration Southeast University Ministry of Education China School of Information and Physical Sciences The University of Newcastle CallaghanNSW2308 Australia

Weakly Supervised Object Detection (WSOD) enables the training of object detection models using only image-level annotations. State-of-the-art WSOD detectors commonly rely on multi-instance learning (MIL) as the backbone of their detectors and assume that the bounding box proposals of an image are independent of each other. However, since such approaches only utilize the highest score proposal and discard the potentially useful information from other proposals, their independent MIL backbone often limits models to salient parts of an object or causes them to detect only one object per class. To solve the above problems, we propose a novel backbone for WSOD based on our tailored Vision Transformer named Weakly Supervised Transformer Detection network (WSTDN). Our algorithm is not only the first to demonstrate that self-attention modules that consider inter-instance relationships are effective backbones for WSOD, but also we introduce a novel bounding box mining method (BBM) integrated with a memory transfer refinement (MTR) procedure to utilize the instance dependencies for facilitating instance refinements. Experimental results on PASCAL VOC2007 and VOC2012 benchmarks demonstrate the effectiveness of our proposed WSTDN and modified instance refinement modules. Copyright © 2023, The Authors. All rights reserved.

关键词： Object detection

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：