检索结果-内蒙古大学图书馆

Proceedings of the 38th International Conference on Neural Information Processing Systems

作者： Zhen Zhao Jingqun Tang Binghong Wu Chunhui Lin Shu Wei Hao Liu Xin Tan Zhizhong Zhang Can Huang Yuan Xie East China Normal University and ByteDance ByteDance East China Normal University East China Normal University and Shanghai Key Laboratory of Computer Software Evaluating and Testing Shanghai China

ISBN: (纸本)9798331314385

In this work, we present TextHarmony, a unified and versatile multimodal generative model proficient in comprehending and generating visual text. Simultaneously generating images and texts typically results in performance degradation due to the inherent inconsistency between vision and language modalities. To overcome this challenge, existing approaches resort to modality-specific data for supervised fine-tuning, necessitating distinct model instances. We propose Slide-LoRA, which dynamically aggregates modality-specific and modality-agnostic LoRA experts, partially decoupling the multimodal generation space. Slide-LoRA harmonizes the generation of vision and language within a singular model instance, thereby facilitating a more unified generative process. Additionally, we develop a high-quality image caption dataset, DetailedTextCaps-100K, synthesized with a sophisticated closed-source MLLM to enhance visual text generation capabilities further. Comprehensive experiments across various benchmarks demonstrate the effectiveness of the proposed approach. Empowered by Slide-LoRA, TextHarmony achieves comparable performance to modality-specific fine-tuning results with only a 2% increase in parameters and shows an average improvement of 2.5% in visual text comprehension tasks and 4.0% in visual text generation tasks. Our work delineates the viability of an integrated approach to multimodal generation within the visual text domain, setting a foundation for subsequent inquiries. Code is available at https://***/bytedance/TextHarmony.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Load Balancing Optimization for Transformer in Distributed Environment 27

Load Balancing Optimization for Transformer in Distributed E...

引用

27th IEEE International Conference on Parallel and Distributed Systems, ICPADS 2021

作者： Ma, Delu Lei, Zhou Chen, Shengbo Wang, Peng School of Computer Engineering and Science Shanghai University Shanghai200444 China Shanghai Key Laboratory of Computer Software Testing and Evaluating Shanghai201112 China

ISBN: (纸本)9781665408783

In recent years, the demand for artificial intelligence applications has increased dramatically. Complex models can promote machine learning to achieve excellent results, but computing efficiency has gradually reached a bottleneck. Therefore, more researchers are exploring the improvement of the efficiency of intelligent computing systems. Distributed machine learning can improve the efficiency of model training and inference, but problems such as communication delay and load imbalance between computing nodes still exist. In the multi-GPU distributed computing environment, this paper takes the vision field algorithm VIT (vision transformer) as the optimization object, which has the advantage of convenient parallel training, and proposes several related solutions. Firstly, the parameter server is used as the system logic architecture and in order to reduce the idleness of the computing devices during the training process, the device working status query mechanism is designed to realize load balancing. Secondly, combined with the pre-trained small VIT algorithm model, semi-asynchronous communication method is proposed to reduce the communication overhead of computing devices and accelerate global convergence. The results of this experiment carried out in the existing distributed environment has demonstrated that compared with the existing synchronization method, the computational efficiency has been improved well under the premise of slightly reducing the accuracy. © 2021 IEEE.

关键词： Computation theory

来源：评论

学校读者我要写书评

暂无评论

Research on stream processing engine and benchmarking framework

Research on stream processing engine and benchmarking framew...

引用

IEEE International Conference on Networking, Sensing and Control

作者： Qionghua Le Mingang Chen Wenjie Chen Shanghai Key Laboratory of Computer Software Testing &Evaluating Shanghai Development Center of Computer Software Technology Shanghai China

ISBN: (纸本)9781665472449

Stream computing engine is an important part of big data system, and benchmarking is one of the main means to measure the engine's performance. In this paper, we compare the differences between two engines, Spark Streaming and Flink, in stream processing technologies. Then the open source benchmarking frameworks supporting stream processing and their respective characteristics are studied, and the HiBench testing framework is selected to test the two stream processing engines. The test results show that Flink is better than Spark Streaming in terms of performance in shuffle, stateful computation and windowed computation.

关键词： Process control Benchmark testing Big Data Sensors Sparks Engines

来源：评论

学校读者我要写书评

暂无评论

Business Scenario Driven Reinforcement Learning testing Method

Business Scenario Driven Reinforcement Learning Testing Meth...

引用

software Engineering, Artificial Intelligence, Networking and Parallel/Distributed Computing (SNPD-Winter), ACIS International Winter Conference on

作者： Lizhi Cai Jin Wang Shanghai Key Laboratory of Computer Software Testing and Evaluating School of Information Science and Engineering East China University of Science and Technology Shanghai China

Reinforcement learning has been successfully applied in software testing, but the existing testing methods cannot perform effective testing according to the characteristics of applications, and using outdated interactive experience during training, resulting in inefficient testing. In this paper, we propose BSDRtesting. Firstly, the demonstration experience of human users is collected according to the functional scenarios and business logic of each application, and combining reinforcement learning and imitation learning to maximize rewards while imitating user behavior, experience replay aims to sample experiences from the agent's self-exploration and expert demonstrations to improve sampling efficiency. At the same time, according to the input rules, the black-box testing method is used to fully test the input events, and finally an experience filtering mechanism is proposed, and the reward value and TD-Error are used as the basis for priority sampling. The experimental results on 10 open source applications show BSDRtesting has achieved significant improvements in code coverage and branch coverage compared with existing methods.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Location-Based Service Recommendation for Cold-Start in Mobile Edge Computing 17th

Location-Based Service Recommendation for Cold-Start in Mobi...

引用

17th IFIP WG 10.3 International Conference on Network and Parallel Computing, NPC 2020

作者： Yu, Mengshan Fan, Guisheng Yu, Huiqun Chen, Liang Department of Computer Science and Engineering East China University of Science and Technology Shanghai China Shanghai Key Laboratory of Computer Software Evaluating and Testing Shanghai China

ISBN: (纸本)9783030794774

With the rapid development of the 5G and Internet of Things (IoT), mobile edge computing has gained considerable popularity in academic and industrial field, which provides physical resources closer to end users. Service recommendation in such a distributed environment is a hot issue. However, the cold-start problem for service recommendation in mobile edge computing is still urgent to be solved. In this paper, we propose a service recommendation method based on collaborative filtering (CF) and user location, by comprehensively considering the characteristic of services at the edge and mobility of users. In detail, we first synthesize the service characteristics of each dimension through multidimensional weighting method. We further introduce the idea of Inverse_CF_Rec to the traditional CF and then predict the lost QoS value to solve the problem of sparse data. Finally, a recommendation algorithm based on predicted QoS value and user geographic location is proposed to recommend appropriate services to users. The experimental results show that our multidimensional inverse similarity recommendation algorithm based on collaborative filtering (MDICF) outperforms Inverse_CF_Rec in terms of the accuracy of recommendation. © 2021, IFIP International Federation for Information Processing.

关键词： Location based services

来源：评论

学校读者我要写书评

暂无评论

Instance-Aware Domain Generalization for Face Anti-Spoofing

arXiv

引用

arXiv 2023年

作者： Zhou, Qianyu Zhang, Ke-Yue Yao, Taiping Lu, Xuequan Yi, Ran Ding, Shouhong Ma, Lizhuang Shanghai Jiao Tong University China Youtu Lab Tencent China Shanghai Key Laboratory of Computer Software Evaluating and Testing China Deakin University Australia

Face anti-spoofing (FAS) based on domain generalization (DG) has been recently studied to improve the generalization on unseen scenarios. Previous methods typically rely on domain labels to align the distribution of each domain for learning domain-invariant representations. However, artificial domain labels are coarse-grained and subjective, which cannot reflect real domain distributions accurately. Besides, such domain-aware methods focus on domain-level alignment, which is not fine-grained enough to ensure that learned representations are insensitive to domain styles. To address these issues, we propose a novel perspective for DG FAS that aligns features on the instance level without the need for domain labels. Specifically, Instance-Aware Domain Generalization framework is proposed to learn the generalizable feature by weakening the features' sensitivity to instance-specific styles. Concretely, we propose Asymmetric Instance Adaptive Whitening to adaptively eliminate the style-sensitive feature correlation, boosting the generalization. Moreover, Dynamic Kernel Generator and Categorical Style Assembly are proposed to first extract the instance-specific features and then generate the style-diversified features with large style shifts, respectively, further facilitating the learning of style-insensitive features. Extensive experiments and analysis demonstrate the superiority of our method over state-of-the-art competitors. Code will be publicly available at this link. Copyright © 2023, The Authors. All rights reserved.

关键词： Machine learning

来源：评论

学校读者我要写书评

暂无评论

Honeytoken-Detector: A Symbolic Execution-Based Honeypot Token Detection Tool

Honeytoken-Detector: A Symbolic Execution-Based Honeypot Tok...

引用

software Engineering, Artificial Intelligence, Networking and Parallel/Distributed Computing (SNPD-Winter), ACIS International Winter Conference on

作者： Yi Liu Lizhi Cai School of Information Science and Engineering East China University of Science and Technology Shanghai China Shanghai Key Laboratory of Computer Testing & Evaluating Shanghai Development Center of Computer Software Technology Shanghai China

ISBN: (数字)9798350345865

ISBN: (纸本)9798350345872

Smart contracts have emerged as one of the most successful applications in the blockchain domain, playing a significant role in various blockchain ecosystems. Inspired by smart contracts, a multitude of cryptographic assets have been created. To standardize these assets, industry standards such as ERC20 (Ethereum Request for Comments 20), ERC721, and ERC1155 have been proposed. In recent years, smart contracts have frequently fallen victim to attacks. Honeypot contracts, disguised as ERC20-compliant tokens, are widely prevalent on the blockchain, enticing victims to make purchases. Such malicious smart contracts exhibiting deceptive behavior are collectively referred to as honeypot tokens. This paper focuses on ERC20-compliant smart contracts and defines six common types of honeypot issues. Building upon advancements in smart contract vulnerability detection, we propose an enhanced symbolic execution-based detection tool called Honeytoken-Detector. We conduct experiments on both contracts known to have similar issues and actual token contracts from the real world. The experimental results demonstrate the effectiveness of our tool in identifying vulnerabilities.

关键词： Industries Source coding Smart contracts Ecosystems Buildings Blockchains Behavioral sciences

来源：评论

学校读者我要写书评

暂无评论

LDE-UNet: A Novel Model for Rapid COVID-19 Diagnosis via CT Image Segmentation

LDE-UNet: A Novel Model for Rapid COVID-19 Diagnosis via CT ...

引用

IEEE International Conference on Bioinformatics and Biomedicine (BIBM)

作者： QiSong Wang JiaRou Wu YingZhuo Wang HuiLi Qu Shanling Nie Hai Yang Department of Computer Science and Engineering East China University of Science and Technology Shanghai China Shanghai Key Laboratory of Computer Software Evaluating and Testing Shanghai China Faculty of Engineering The University of Sydney Australia

The COVID-19 pandemic has had a profound impact on human society. It has highlighted the need for faster diagnostic methods. Research has shown that combining semantic segmentation with traditional medical approaches can significantly accelerate the process. To address this, leveraging COVID-19 CT images, our team designs a revolutionary semantic segmentation model called Level of Detail Enhancement U-Net (LDE-UNet), which shows the lesion area on CT images. By introducing the LDE block, the model has the unique advantage of overcoming the loss of data details during the downsampling process by emphasizing and transmitting details at the same level. Our SOTA model outperforms the second-best model by at least 0.7% in the most critical indicator precision. Compared with other models, LDE-UNet’s strong reliability determines its ability to be used in the medical field to accelerate the localization and division of lesion areas on CT images by professional doctors, thus completing patient diagnosis faster. In addition, we also propose a standardized method for processing medical images.

关键词：

来源：评论

学校读者我要写书评

暂无评论

CHESS: Learning User Correlation from Semantic-related Cascades for Diffusion Prediction 9

CHESS: Learning User Correlation from Semantic-related Casca...

引用

9th IEEE Smart World Congress, SWC 2023

作者： Li, Huacheng Xia, Chunhe Wang, Tianbo Zhao, Haopeng Beihang University Key Laboratory of Beijing Network Technology Beijing China Guangxi Collab. Innov. Ctr. of Multi-Src. Info. Intgn. and Intelligent Proc. Guangxi Norm. Univ. Guilin China Beihang University School of Cyber Science and Technology Beijing China Shanghai Key Laboratory of Computer Software Evaluating and Testing Shanghai China

ISBN: (纸本)9798350319804

Understanding the information diffusion trajectory in online social networks has become an ever-growing prominent research direction in the information era. Users interact with social information they are interested in, and cascades record diffusion trajectories of massive social information. It implies two characteristics of diffusion cascades: topic sharing among different cascades and infected user correlation within the same cascades. However, previous works ignore these two characteristics and primarily focus on whether each infected user is correctly predicted individually, which leads to learning biased diffusion models under the data-driven mode. To address this problem, we propose a novel architecture to exploit such latent user associations among topic-related cascades. First, we leverage the hypergraph and heterogeneous graph to exploit infected user correlations jointly and generate user feature vectors. Second, we devise an enhanced cascade similarity loss function to evaluate prediction results from delay (horizontal) and rank (vertical) scopes. Furthermore, we introduce the extra predictor to alleviate the interference with the primary model. Experiments on four public datasets validate the effectiveness of our proposed method and significantly outperform existing models on hits@κ and map@κ indicators. © 2023 IEEE.

关键词： heterogeneous hyper graph information diffusion prediction loss function social network analysis

来源：评论

学校读者我要写书评

暂无评论

CIXG: A Comprehensive Approach to Driver Gene Identification and Causal Interpretation

CIXG: A Comprehensive Approach to Driver Gene Identification...

引用

IEEE International Conference on Bioinformatics and Biomedicine (BIBM)

作者： Yawen Liu Yiwen Chen Shanling Nie Hai Yang Department of Computer Science and Engineering East China University of Science and Technology Shanghai China Shanghai Key Laboratory of Computer Software Evaluating and Testing Shanghai China Center for Continuing and Lifelong Education National University of Singapore Faculty of Engineering The University of Sydney Australia

ISBN: (数字)9798350386226

ISBN: (纸本)9798350386233

With the ongoing advancements in science and technology and the increasing research focus on cancer-related issues, there has been a proliferation of omics-related resources for in-depth analysis and exploration. This burgeoning volume and complexity of biological data have fostered the integration of machine-learning techniques into biology. As a result, numerous machine-learning strategies have been established to identify driver mutations. Yet, many of these strategies produce complex models, complicating comprehension and thereby clouding the impact of input features on the resulting predictions. Our analysis presented the CIXG framework, which integrates a driver gene prediction module using XGBoost with a causality interpretation module anchored on CXPlain. This architecture enables quantifying each input feature’s contribution to the prediction outcome and ensures precise predictions of driver genes. When benchmarked against the state-of-the-art (SOTA) method, CIXG demonstrated superior accuracy in pinpointing driver genes across pan-cancer studies and within the 32 specific cancer types. Importantly, our results underscored that mutation features chiefly influence CIXG’s predictive prowess, with additional support from other omics features.

关键词： Accuracy Biological system modeling Cause effect analysis Machine learning Predictive models Benchmark testing Complexity theory Bioinformatics Cancer

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：