检索结果-南通市图书馆

Social conformity creates consensus and strong diversity of Hegselmann-Krause opinion dynamics

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

science China(Information sciences) 2022年第2期65卷 251-253页

作者： Chun CHENG Changbin YU school of computer science Fudan University school of Engineering Westlake University

Dear editor,Over the past decade,"opinion dynamics"of social networks have obtained considerable attention from control theory,sociology,and physics[1–3].One of the most commonly studied models is the Hegselmann-Krause (H-K) model[4],which is boundedly confident,so that each agent only updates his/her opinion by averaging all the opinions of ***,the model has been criticized for a considerable time because limiting behaviour such as weak diversity has been pointed out in[5].

关键词： opinion dynamics Hegselmann-Krause model social network conformity multi-agent systems

An Efficient Approach of Processing Multiple Continuous Queries

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

Journal of computer science & Technology 2016年第6期31卷 1212-1227页

作者： Wen Liu Yan-Ming Shen Peng Wang school of computer science and Technology Dalian University of Technology Dalian 116024 China school of computer science Fudan University Shanghai 201203 China

As stream data is being more frequently collected and analyzed, stream processing systems are faced with more design challenges. One challenge is to perform continuous window aggregation, which involves intensive computation. When there are a large number of aggregation queries, the system may suffer from scalability problems. The queries are usually similar and only differ in window specifications. In this paper, we propose collaborative aggregation which promotes aggregate sharing among the windows so that repeated aggregate operations can be avoided. Different from the previous approaches in which the aggregate sharing is restricted by the window pace, we generalize the aggregation over multiple values as a series of reductions. Therefore, the results generated by each reduction step can be shared. The sharing process is formalized in the feed semantics and we present the compose-and-declare framework to determine the data sharing logic at a very low cost. Experimental results show that our approach offers an order of magnitude performance improvement to the state-of-the-art results and has a small memory footprint.

关键词： data stream streams aggregation query sharing continuous query

Generative API usage code recommendation with parameter concretization

维普期刊数据库评论

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

science China(Information sciences) 2019年第9期62卷 51-72页

作者： Chi CHEN Xin PENG Jun SUN Zhenchang XING Xin WANG Yifan ZHAO Hairui ZHANG Wenyun ZHAO school of computer science Fudan University Shanghai Key Laboratory of Data science Fudan University Pillar of Information System Technology and Design Singapore University of Technology and Design Research school of computer science Australian National University

Many programming languages and development frameworks have extensive libraries(e.g., JDK and Android libraries) that ease the task of software engineering if used effectively. With numerous library classes and sometimes intricate API(application programming interface) usage constraints, programmers often have difficulty remembering the library APIs and/or using them correctly. This study addresses this problem by developing an engine called DeepAPIRec, which automatically recommends the API usage *** to the existing proposals, our approach distinguishes itself in two ways. First, it is based on a tree-based long short-term memory(LSTM) neural network inspired by recent developments in the machinelearning community. A tree-based LSTM neural network allows us to model and reason about variable-length,preceding and succeeding code contexts, and to make precise predictions. Second, we apply data-flow analysis to generate concrete parameters for the API usage code, which not only allows us to generate complete code recommendations but also improves the accuracy of the learning results according to the tree-based LSTM neural network. Our approach has been implemented for supporting Java programs. Our experimental studies on the JDK library show that at statement-level recommendations, DeepAPIRec can achieve a top-1 accuracy of about 37% and a top-5 accuracy of about 64%, which are significantly better than the existing approaches. Our user study further confirms that DeepAPIRec can help developers to complete a segment of code faster and more accurately as compared to IntelliJ IDEA.

关键词： code recommendation API deep learning data dependency parameter concretization

Floating Waste Discovery by Request via Object-Centric Learning

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

computers, Materials & Continua 2024年第7期80卷 1407-1424页

作者： Bingfei Fu school of computer science Fudan UniversityShanghai200438China

Discovering floating wastes,especially bottles on water,is a crucial research problem in environmental ***,real-world applications often face challenges such as interference from irrelevant objects and the high cost associated with data ***,devising algorithms capable of accurately localizing specific objects within a scene in scenarios where annotated data is limited remains a formidable *** solve this problem,this paper proposes an object discovery by request problem setting and a corresponding algorithmic *** proposed problem setting aims to identify specified objects in scenes,and the associated algorithmic framework comprises pseudo data generation and object discovery by request ***-data generation generates images resembling natural scenes through various data augmentation rules,using a small number of object samples and scene *** network structure of object discovery by request utilizes the pre-trained Vision Transformer(ViT)model as the backbone,employs object-centric methods to learn the latent representations of foreground objects,and applies patch-level reconstruction constraints to the *** the validation phase,we use the generated pseudo datasets as training sets and evaluate the performance of our model on the original test *** have proved that our method achieves state-of-the-art performance on Unmanned Aerial Vehicles-Bottle Detection(UAV-BD)dataset and self-constructed dataset Bottle,especially in multi-object scenarios.

关键词： Unsupervised object discovery object-centric learning pseudo data generation real-world object discovery by request

Adaptive Backdoor Attack against Deep Neural Networks

维普期刊数据库评论

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

computer Modeling in Engineering & sciences 2023年第9期136卷 2617-2633页

作者： Honglu He Zhiying Zhu Xinpeng Zhang school of computer science Fudan UniversityShanghai200433China

In recent years,the number of parameters of deep neural networks(DNNs)has been increasing *** training of DNNs is typically *** a result,many users leverage cloud computing and outsource their training *** computation results in a potential risk called backdoor attack,in which a welltrained DNN would performabnormally on inputs with a certain *** attacks can also be classified as attacks that exploit fake ***,most backdoor attacks design a uniformtrigger for all images,which can be easilydetectedand *** this paper,we propose a novel adaptivebackdoor *** overcome this defect and design a generator to assign a unique trigger for each image depending on its *** achieve this goal,we use a texture complexitymetric to create a specialmask for eachimage,which forces the trigger tobe embedded into the rich texture *** trigger is distributed in texture regions,which makes it invisible to *** the stealthiness of triggers,we limit the range of modification of backdoor models to evade *** show that our method is efficient in multiple datasets,and traditional detectors cannot reveal the existence of a backdoor.

关键词： Backdoor attack AI security DNN

CPT: a pre-trained unbalanced transformer for both Chinese language understanding and generation

维普期刊数据库评论

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

science China(Information sciences) 2024年第5期67卷 43-55页

作者： Yunfan SHAO Zhichao GENG Yitao LIU Junqi DAI Hang YAN Fei YANG Zhe LI Hujun BAO Xipeng QIU school of computer science Fudan University Shanghai Key Laboratory of Intelligent Information Processing Fudan University Zhejiang Lab

In this paper, we take the advantage of previous pre-trained models(PTMs) and propose a novel Chinese pre-trained unbalanced transformer(CPT). Different from previous Chinese PTMs, CPT is designed to utilize the shared knowledge between natural language understanding(NLU) and natural language generation(NLG) to boost the performance. CPT consists of three parts: a shared encoder, an understanding decoder, and a generation decoder. Two specific decoders with a shared encoder are pretrained with masked language modeling(MLM) and denoising auto-encoding(DAE) tasks, *** the partially shared architecture and multi-task pre-training, CPT can(1) learn specific knowledge of both NLU or NLG tasks with two decoders and(2) be fine-tuned flexibly that fully exploits the potential of the model. Moreover, the unbalanced transformer saves the computational and storage cost, which makes CPT competitive and greatly accelerates the inference of text generation. Experimental results on a wide range of Chinese NLU and NLG tasks show the effectiveness of CPT.

关键词： pre-trained model transformer language model generation unified model

同方期刊数据库

How security bugs are fixed and what can be improved: an empirical study with Mozilla

在线全文

学校读者我要写书评

暂无评论

science China(Information sciences) 2019年第1期62卷 214-216页

作者： Xiaobing SUN Xin PENG Kai ZHANG Yang LIU Yuanfang CAI school of computer science Fudan University Shanghai Key Laboratory of Data science Fudan University school of Information Engineering Yangzhou University school of computer science and Engineering Nanyang Technological University Department of computer science Drexel University

Dear editor,A bug is regarded as security related when it creates vulnerability in the software, which the malicious attackers could exploit to attack the system [1]. Security bugs widely exist in released software, and are serious threats to organizations,which can cause serious monetary loss and reputation damage. Therefore, addressing security bugs is a top priority in software maintenance [1].

关键词： How security bugs are fixed and what can be improved an empirical study with Mozilla

AI-boosted software automation:learning from human pair programmers

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

science China(Information sciences) 2019年第10期62卷 50-52页

作者： Xin PENG Zhenchang XING Jun SUN school of computer science Fudan University Shanghai Key Laboratory of Data science Fudan University Research school of computer science Australian National University school of Information Systems Singapore Management University

Dear editor,Software automation [1] aims to automatically generate computer programs from formal or informal requirement descriptions. It covers a variety of transformations of different spans, including generating programs from natural-language requirements, requirements specifications, or design specifications. Traditionally software automation is achieved through logical reasoning and rule-based transformation [1]. Although the transformation

关键词： AI-boosted software automation:learning from human pair programmers API pro

Miyaoka-type inequalities for terminal threefolds with nef anti-canonical divisors

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

science China Mathematics 2025年第1期 1-18页

作者： Masataka Iwai Chen Jiang Haidong Liu Department of Mathematics Graduate School of ScienceOsaka University Shanghai Center for Mathematical sciences & school of Mathematical sciences Fudan University Department of mathematics Sun Yat-sen University

In this paper,we study Miyaoka-type inequalities on Chern classes of terminal projective 3-folds with nef anti-canonical *** X be a terminal projective 3-fold such that-KXis *** show that if c1（X） · c2（X）≠0,then c1（X） · c2（X） ≥ 1/252;if further X is not rationally connected,then c1（X） · c2（X） ≥ 4/5 and this inequality is *** order to prove this,we give a partial classification of such varieties along with many *** also study the nonvanishing of c1（X）dim X-2· c2（X） for terminal weak Fano varieties and prove a Miyaoka-Kawamata-type inequality.

关键词：