首页期刊导航|中国科学：信息科学（英文版）
期刊信息/Journal information
中国科学：信息科学（英文版）
中国科学：信息科学（英文版）

周光召

月刊

1674-733X

informatics@scichina.org

010-64015683

100717

北京东黄城根北街16号

中国科学：信息科学（英文版）/Journal Science China Information SciencesCSCDCSTPCDEISCI
查看更多>>《中国科学》是中国科学院主办、中国科学杂志社出版的自然科学专业性学术刊物。《中国科学》任务是反映中国自然科学各学科中的最新科研成果，以促进国内外的学术交流。《中国科学》以论文形式报道中国基础研究和应用研究方面具有创造性的、高水平的和有重要意义的科研成果。在国际学术界，《中国科学》作为代表中国最高水平的学术刊物也受到高度重视。国际上最具有权威的检索刊物SCI，多年来一直收录《中国科学》的论文。1999年《中国科学》夺得国家期刊奖的第一名。
正式出版
收录年代

    Recent progress in single crystal perovskite X-ray detectors

    Xiao LIUJun RENYu-Ang CHENXiangshun GENG...
    1-15页
    查看更多>>摘要：Perovskites have attracted extensive attention as radiation detection material due to their long carrier diffusion length and lifetime,high absorption coefficient,and flexible manufacturing process.Com-pared with polycrystalline structures,single crystal perovskites improve the performance of optoelectronic devices due to their low defect state density,better photoelectric characteristics,and chemical stability.Herein,we review the recent progress of single crystal perovskite X-ray detectors.First,we briefly intro-duced the basic concepts,detection mechanisms,figure of merits of perovskite X-ray detectors,and the preparation methods of single crystal perovskites.Then,we summarized the significant advancements in single crystal perovskite X-ray detectors in recent times.Finally,we discussed the critical challenges and some practicable solutions for developing high-performance X-ray detectors.
      原文链接:
    • 万方数据
    • 维普

    Model architecture level privacy leakage in neural networks

    Yan LIHongyang YANTeng HUANGZijie PAN...
    16-28页
    查看更多>>摘要：Privacy leakage is one of the most critical issues in machine learning and has attracted growing interest for tasks such as demonstrating potential threats in model attacks and creating model defenses.In recent years,numerous studies have revealed various privacy leakage risks(e.g.,data reconstruction at-tack,membership inference attack,backdoor attack,and adversarial attack)and several targeted defense approaches(e.g.,data denoising,differential privacy,and data encryption).However,existing solutions gen-erally focus on model parameter levels to disclose(or repair)privacy threats during the model training and/or model interference process,which are rarely applied at the model architecture level.Thus,in this paper,we aim to exploit the potential privacy leakage at the model architecture level through a pioneer study on neural architecture search(NAS)paradigms which serves as a powerful tool to automate a neural network design.By investigating the NAS procedure,we discover two attack threats in the model architecture level called the architectural dataset reconstruction attack and the architectural membership inference attack.Our the-oretical analysis and experimental evaluation reveal that an attacker may leverage the output architecture of an ongoing NAS paradigm to reconstruct its original training set,or accurately infer the memberships of its training set simply from the model architecture.In this work,we also propose several defense approaches related to these model architecture attacks.We hope our work can highlight the need for greater attention to privacy protection in model architecture levels(e.g.,NAS paradigms).
      原文链接:
    • 万方数据
    • 维普

    Practical cloud storage auditing using serverless computing

    Fei CHENJianquan CAITao XIANGXiaofeng LIAO...
    29-43页
    查看更多>>摘要：Cloud storage auditing research is dedicated to solving the data integrity problem of outsourced storage on the cloud.In recent years,researchers have proposed various cloud storage auditing schemes using different techniques.While these studies are elegant in theory,they assume an ideal cloud storage model;that is,they assume that the cloud provides the storage and compute interfaces as required by the proposed schemes.However,this does not hold for mainstream cloud storage systems because these systems only provide read and write interfaces but not the compute interface.To bridge this gap,this work proposes a serverless computing-based cloud storage auditing system for existing mainstream cloud object storage.The proposed system leverages existing cloud storage auditing schemes as a basic building block and makes two adaptations.One is that we use the read interface of cloud object storage to support block data requests in a traditional cloud storage auditing scheme.Another is that we employ the serverless computing paradigm to support block data computation as traditionally required.Leveraging the characteristics of serverless computing,the proposed system realizes economical,pay-as-you-go cloud storage auditing.The proposed system also supports mainstream cloud storage upper layer applications(e.g.,file preview)by not modifying the data formats when embedding authentication tags for later auditing.We prototyped and open-sourced the proposed system to a mainstream cloud service,i.e.,Tencent Cloud.Experimental results show that the proposed system is efficient and promising for practical use.For 40 GB of data,auditing takes approximately 98 s using serverless computation.The economic cost is 120.48 CNY per year,of which serverless computing only accounts for 46％.In contrast,no existing studies reported cloud storage auditing results for real-world cloud services.
      原文链接:
    • 万方数据
    • 维普

    Multi-instance partial-label learning:towards exploiting dual inexact supervision

    Wei TANGWeijia ZHANGMin-Ling ZHANG
    44-57页
    查看更多>>摘要：Weakly supervised machine learning algorithms are able to learn from ambiguous samples or labels,e.g.,multi-instance learning or partial-label learning.However,in some real-world tasks,each training sample is associated with not only multiple instances but also a candidate label set that contains one ground-truth label and some false positive labels.Specifically,at least one instance pertains to the ground-truth label while no instance belongs to the false positive labels.In this paper,we formalize such problems as multi-instance partial-label learning(MIPL).Existing multi-instance learning algorithms and partial-label learning algorithms are suboptimal for solving MIPL problems since the former fails to disambiguate a candidate label set,and the latter cannot handle a multi-instance bag.To address these issues,a tailored algorithm named MIPLGP,i.e.,multi-instance partial-label learning with Gaussian processes,is proposed.MIPLGP first assigns each instance with a candidate label set in an augmented label space,then transforms the candidate label set into a logarithmic space to yield the disambiguated and continuous labels via an exclusive disambiguation strategy,and last induces a model based on the Gaussian processes.Experimental results on various datasets validate that MIPLGP is superior to well-established multi-instance learning and partial-label learning algorithms for solving MIPL problems.
      原文链接:
    • 万方数据
    • 维普

    Mitigate noisy data for smart IoT via GAN based machine unlearning

    Zhuo MAYilong YANGYang LIUXinjing LIU...
    58-74页
    查看更多>>摘要：With the development of IoT applications,machine learning dramatically improves the utility of variable IoT systems such as autonomous driving.Although the pretrain-finetune framework can cope well with data heterogeneity in complex IoT scenarios,the data collected by sensors often contain unexpected noisy data,e.g.,out-of-distribution(OOD)data,which leads to the reduced performance of fine-tuned models.To resolve the problem,this paper proposes MuGAN,a method that can mitigate the side-effect of OOD data via the generative adversarial network(GAN)-based machine unlearning.MuGAN follows a straightforward but effective idea to mitigate the performance loss caused by OOD data,i.e.,"flashbacking"the model to the condition where OOD data are excluded from model training.To achieve the goal,we design an adversarial game,where a discriminator is trained to identify whether a sample belongs to the training set by observing the confidence score.Meanwhile,a generator(i.e.,the target model)is updated to fool the discriminator into believing that the OOD data are not included in the training set but others do.The experimental results show that benefiting from the high unlearning rate(more than 90％)and retention rate(99％),MuGAN succeeds in lowering the model performance degradation caused by OOD data from 5.88％to 0.8％.
      原文链接:
    • 万方数据
    • 维普

    Unbalanced private set intersection with linear communication complexity

    Quanyu ZHAOBingbing JIANGYuan ZHANGHeng WANG...
    75-89页
    查看更多>>摘要：The private set intersection(PSI)protocol allows two parties holding a set of integers to compute the intersection of their sets without revealing any additional information to each other.The unbalanced PSI schemes consider a specific setting where a client holds a small set of the size n and a server holds a much larger set of the size m(n<<m).The communication overhead of state-of-the-art balanced PSI schemes is O(m+n)and the unbalanced PSI schemes are O(nlogm).In this paper,we propose a novel secure unbalanced PSI protocol based on a hash proof system.The communication complexity of our protocol grows only linearly with the size of the small set.In other words,our protocol achieves communication overhead of O(n).We test the performance on a personal computer(PC)machine with a local area network(LAN)setting for the network.The experimental results demonstrate that the client only takes 2.01 s of online computation,4.27 MB of round trip communication to intersect 1600 pieces of 32-bit integers with 220 pieces of 32-bit integers with the security parameter λ=512.Our protocol is efficient and can be applied to resource-constrained devices,such as cell phones.
      原文链接:
    • 万方数据
    • 维普

    Distinct but correct:generating diversified and entity-revised medical response

    Bin LIBin SUNShutao LIEncheng CHEN...
    90-109页
    查看更多>>摘要：Medical dialogue generation(MDG)is applied for building medical dialogue systems for intelli-gent consultation.Such systems can communicate with patients in real time,thereby improving the efficiency of clinical diagnosis.However,predicting correct entities and correctly generating distinct responses remain a great challenge.Inspired by actual doctors'responses to patients,we consider MDG a two-stage task:entity prediction and dialogue generation.For entity prediction,we design an ent-mac post pre-training strategy by leveraging external medical entity knowledge to enhance the pre-trained model.For dialogue genera-tion,we propose an entity-aware fusion MDG method in which predicted entities are integrated into the dialogue generation model through different encoding fusion mechanisms,using information from different sources.Because the diverse beam search algorithm can produce responses with entities that deviate from the predicted entities,an entity-revised diverse beam search is proposed to correct the entities entailed in the generated responses and make the generated responses more distinct.The experimental results on the China Conference on Knowledge Graph and Semantic Computing 2021(A/B tests)and the International Confer-ence on Learning Representations 2021(online test)datasets show that the proposed method outperforms several state-of-the-art methods,which demonstrates its practicability and effectiveness.
      原文链接:
    • 万方数据
    • 维普

    A FAS approach for stabilization of generalized chained forms:part 2.Continuous control laws

    Guang-Ren DUAN
    110-131页
    查看更多>>摘要：In this paper,continuous time-varying stabilizing controllers for the type of general nonholonomic systems proposed and treated in part 1 are designed using the fully actuated system(FAS)approach.The key step is to differentiate the first scalar equation,and by control of the obtained second-order scalar system,a proportional plus integral feedback form for the first control variable is obtained.With the solution to this designed second-order scalar system,the rest equations in the nonholonomic system form an independent time-varying subsystem which is then handled by the FAS approach.The overall designed controller contains an almost arbitrarily chosen design parameter,and is proven to guarantee the uniformly and globally exponential stability of the closed-loop system.The proposed approach is simple and effective,and is demonstrated with a practical example of ship control.
      原文链接:
    • 万方数据
    • 维普

    Constrained reinforcement learning with statewise projection:a control barrier function approach

    Xinze JINKuo LIQingshan JIA
    132-150页
    查看更多>>摘要：Safety is a critical issue for reinforcement learning(RL),as it may be risky for some actual appli-cations if the learning process involves unsafe exploration.Instead of formulating constraints as expectation-based in constrained RL,considering statewise safety in constrained RL is more meaningful.This work aims to address the issue of safe projection in RL by introducing a control barrier function that inherently learns a safe policy through a set certificate.We seek to analyze some theoretical properties of safe projection in the learning process,including convergence and performance bound,and extend the discussion into ensembles and guided controllers.Moreover,we approach analytical solutions for deterministic and stochastic system dynamics.Experimental results in different tasks show that the proposed method achieves better effects in terms of both performance and safety.
      原文链接:
    • 万方数据
    • 维普

    Segment-wise learning control for trajectory tracking of robot manipulators under iteration-dependent periods

    Fan ZHANGDeyuan MENGKaiquan CAI
    151-162页
    查看更多>>摘要：This paper is concerned with the amplitude boundedness problem of adaptive iterative learning control(AILC)for robot manipulators operating with iteration-dependent periods.By introducing virtual memory slots for storing historical data,a practical AILC method is proposed to achieve the segment-wise learning.This method requires less memory storage for historical information of previous iterations,especially in comparison with that of the conventional AILC methods using point-wise learning strategies.It is shown that not only the energy boundedness but also the amplitude boundedness of estimates and inputs of practical AILC can be guaranteed.Moreover,the practical AILC method can achieve the perfect tracking objective regardless of iteration-dependent periods when the robot manipulators have a persistent full learning property.In addition,a solution to the visual manipulator platform is provided and deployed based on Coppeliasim and Matlab,which helps to show the amplitude boundedness of learning results and the perfect tracking performances of the proposed practical AILC method for robot manipulators.
      原文链接:
    • 万方数据
    • 维普