IEEE Transactions on Services Computing最新文献

{"title":"Secure Cooperative Offloading Across Domains With DIB-Enhanced Multi-Agent RL and Credence Mutual Information","authors":"Kexin Li;Changjian Xu;Hongwei Zhu;Wen Yin;Yu Liu;Yu Peng;Yong Wang;Xingwei Wang;Schahram Dustdar","doi":"10.1109/TSC.2026.3725533","DOIUrl":"10.1109/TSC.2026.3725533","url":null,"abstract":"Multi-access Edge Computing (MEC) supports latency-critical applications while coping with heterogeneous edge resources and stringent privacy requirements. In practice, task offloading is hindered by security risks in cross-domain transmission and by insufficient coordination among decentralized user devices. To address these challenges, we propose <bold>MITRA-DIB</b>, a distributed computation-offloading framework based on multi-agent reinforcement learning (MARL) for <italic>Domain-Secure MEC (DS-MEC)</i>. MITRA-DIB introduces <bold>DIB (Decision Information Bottleneck)</b> to learn task-adaptive representations that preserve essential decision-making features while suppressing decision-irrelevant perturbations, thereby mitigating spurious correlations and reducing unnecessary inter-agent coupling. To promote effective cooperation under centralized training with decentralized execution (CTDE), we further design a <italic>Credence Mutual Information (CMI)</i> auxiliary reward computed via neural Mutual Information estimators, and we incorporate quality-aware experience shaping with positive/negative buffers to selectively amplify beneficial interactions while attenuating detrimental ones. Moreover, we enforce a domain-security rule that constrains sensitive tasks to within-domain execution and secures cross-domain transmission of normal tasks through lightweight cryptography primitives. Extensive experiments show that MITRA-DIB consistently outperforms state-of-the-art baselines, improving the task success rate by up to 30.3% in complex scenarios while achieving lower system costs and faster convergence.","PeriodicalId":13255,"journal":{"name":"IEEE Transactions on Services Computing","volume":"19 5","pages":"3607-3620"},"PeriodicalIF":6.2,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148768152","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"RACE: Risk-Aware Cloud-Edge Orchestration With Online Adaptive Calibration for Trustworthy LLM Services","authors":"Zhihao Wang;Min Ren;Xia Liang;Gaoyong Li;Ding Ding;Mauro Conti","doi":"10.1109/TSC.2026.3726297","DOIUrl":"10.1109/TSC.2026.3726297","url":null,"abstract":"Edge-deployed large language models (LLMs) exhibit substantially higher hallucination rates on knowledge-intensive queries than their cloud counterparts, yet existing edge-cloud scheduling methods route requests based solely on latency or load balancing, lacking any awareness of—or guarantees on—semantic correctness. Reliable routing hinges on quantifying semantic risk and guaranteeing error rates. Semantic entropy quantifies semantic risk, and online adaptive calibration—inspired by conformal prediction—offers error rate control under distribution shift. However, conventional semantic entropy requires multiple sampling passes (<inline-formula><tex-math>$approx$</tex-math></inline-formula>6 s), infeasible on edge devices; static thresholds cannot adapt to shift, so error rates far exceed the target. We propose RACE (Risk-Aware Cloud-Edge Orchestration), the first framework to integrate online adaptive threshold calibration with semantic risk budgeting, providing provable (<inline-formula><tex-math>$T$</tex-math></inline-formula>-independent) regret and finite-sample coverage guarantees. RACE introduces three innovations: a Semantic Entropy Probe (SEP) predicting semantic uncertainty from prompt-only hidden states in a single forward pass (<inline-formula><tex-math>$&lt; $</tex-math></inline-formula>1 ms), enabling real-time hallucination detection ; an online calibrator fusing risk control with online convex optimization to adjust thresholds under shift; and a semantic risk budget grounded in Lagrangian duality that bounds the cumulative hallucination rate below a user-specified tolerance <inline-formula><tex-math>$alpha$</tex-math></inline-formula>. On MMLU and TriviaQA with GPT-5 as the cloud worker and Phi-3-medium as the edge model, RACE achieves a cloud invocation rate of only 4.05–6.25% while keeping edge error below <inline-formula><tex-math>$alpha {=}30%$</tex-math></inline-formula>—the lowest among risk-controlled baselines (Static-CP, ACI, L-ARC); the weaker Mistral-7B exposes the feasibility boundary of <inline-formula><tex-math>$alpha$</tex-math></inline-formula> and the role of the adaptation budget <inline-formula><tex-math>$rho$</tex-math></inline-formula>.","PeriodicalId":13255,"journal":{"name":"IEEE Transactions on Services Computing","volume":"19 5","pages":"3581-3593"},"PeriodicalIF":6.2,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148806378","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Spatio-Temporal Provisioning in Multi-UAV MEC With Pricing Games and Evolutionary Optimization","authors":"Yongxiang Kuang;Bin Jiang;Yongxin Liu;Houbing Herbert Song","doi":"10.1109/TSC.2026.3729801","DOIUrl":"https://doi.org/10.1109/TSC.2026.3729801","url":null,"abstract":"Spatio-temporal volatility of mobile demand and the surge of computation-intensive, delay-sensitive applications pose significant challenges to cost-aware Mobile Edge Computing (MEC) deployments. This paper investigates a multi-UAV-assisted MEC framework and proposes a hierarchical three-time-scale resource provisioning and allocation scheme. At the large time scale, we optimize long-horizon resource provisioning decisions to jointly minimize service degradation and resource cost. At the medium time scale, a spatio-temporal fluctuation model is established for urban users and a Future-Aware UAV Strategy (FAUS) is introduced to proactively dispatch UAV-mounted servers toward predicted hotspot regions. Meanwhile, we design a semi-competitive and semi-cooperative Stackelberg pricing game between stationary edge servers and UAV clusters to ensure individual rationality and induce effective workload splitting. At the small time scale, we implement a lightweight per-slot routine for price-aware offloadingand KKT-based resource allocation for efficient execution. To solve the resulting expensive black-box provisioning optimization, we develop an Environment Switching Evolutionary Optimization (ESEO) algorithm to accelerate convergence and generate diverse Pareto-optimal solutions, enabling flexible choices such as resource-saving MEC and performance-oriented MEC. Extensive evaluations based on a real urban mobility dataset demonstrate that the proposed scheme outperforms benchmark methods in reducing offload failures and delay while providing flexible cost–performance trade-offs.","PeriodicalId":13255,"journal":{"name":"IEEE Transactions on Services Computing","volume":"19 5","pages":"3634-3647"},"PeriodicalIF":6.2,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149133488","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"EcoVMS: LLM-Driven Reinforcement Learning for End-to-End Energy-Optimized VM Scheduling","authors":"Xiaojun Xu;Jingjing Hu;Jialing He;Xiuqi Yang;Zeyang Liu;Zhuo Chen;Jiacheng Wang","doi":"10.1109/TSC.2026.3720939","DOIUrl":"https://doi.org/10.1109/TSC.2026.3720939","url":null,"abstract":"The rapid expansion of the cloud computing ecosystem necessitates scalable and energy-efficient resource management. However, achieving optimal energy efficiency is severely constrained by the complex interplay between computational workloads and thermal dynamics. Traditional data center (DC) scheduling methods typically rely on computationally expensive mechanistic models, such as computational fluid dynamics (CFD), or inefficient algorithm search, which frequently yield suboptimal solutions and struggle to adapt to real-time environments. To address these challenges, we propose an ecology-conservation-optimization virtual machine scheduling (EcoVMS) framework, a data-driven and end-to-end framework for compute-energy co-optimization. At its core, EcoVMS utilizes a latent reward generation enhanced proximal policy optimization (LARGE-PPO) algorithm, which utilizes large language models (LLMs) to analyze runtime telemetry—including central processing unit (CPU) utilization, thermal boundaries, and power consumption, and then utilizes large-scale semantic priors to generate potential reward factors. This mechanism resolves the sparse reward and credit assignment problems inherent in deep reinforcement learning without requiring complex physical modeling. Extensive simulations demonstrate that EcoVMS significantly outperforms baseline methods across diverse workload scenarios. Under a typical 75% workload, EcoVMS achieved a 12.4% improvement in significant energy savings, improving computing resource utilization while reducing hotspots and lowering service latency, resulting in guaranteed quality of service (QoS).","PeriodicalId":13255,"journal":{"name":"IEEE Transactions on Services Computing","volume":"19 5","pages":"3447-3460"},"PeriodicalIF":6.2,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149133489","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"HSSCOA: Constant-Round Secure Collaborative Analytics on Secret-Shared Database","authors":"Jiehang Zhuang;Zoe Lin Jiang;Jianting Ning;Peng Yang;Siu-Ming Yiu;Chuanyi Liu;Junbin Fang;Xuan Wang","doi":"10.1109/TSC.2026.3685314","DOIUrl":"10.1109/TSC.2026.3685314","url":null,"abstract":"Secure collaborative analytics enables multiple data owners to contribute their data to perform query analytics without learning each other’s data. Although current systems are well-optimized for horizontal scenarios, they are not as effective in vertical scenarios. The join operator is the key performance bottleneck of vertically collaborative data analysis. The existing secure nested-loop two-way join approach results in <inline-formula><tex-math>$O(n^{2})$</tex-math></inline-formula> computation and communication complexity, where <inline-formula><tex-math>$n$</tex-math></inline-formula> represents table size. In this paper, we propose HSSCOA, a secure collaborative analytics system to execute join-aggregation queries on secret-shared database. particularly, we propose a simplified homomorphic secret sharing (HSS) supporting just single multiplication, called one-time multiplication HSS. Combined with secret-shared sorting protocol, we then design a secure and efficient two-way join protocol that achieves constant rounds of interaction and <inline-formula><tex-math>$O(n log n)$</tex-math></inline-formula> communication. We further extend our framework to support multi-way joins and various join variants. Comprehensive experiments implemented on MP-SPDZ demonstrate that HSSCOA significantly outperforms existing sort-based protocols in WAN settings due to its constant-round design.","PeriodicalId":13255,"journal":{"name":"IEEE Transactions on Services Computing","volume":"19 5","pages":"3184-3197"},"PeriodicalIF":6.2,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147708688","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Discovering Work Specialization Through Process Mining","authors":"Jing Yang;Chun Ouyang","doi":"10.1109/TSC.2026.3717882","DOIUrl":"10.1109/TSC.2026.3717882","url":null,"abstract":"The success of an organization depends on how well its workforce is mobilized and managed. Many organizations are taking up workforce analytics to support critical decision-making on their workforce. Process mining can empower workforce analytics, using event logs to extract accurate insights into employee performance and collaboration. To achieve precise analysis outcomes, it is imperative to consider work specialization, i.e., division of labor among human resources. In process mining, execution contexts can be used to characterize how specialization manifests across multiple process dimensions and therefore enhance resource-oriented analyses, such as organizational model mining and analysis of resource group work profiles. However, discovering execution contexts from event logs with minimal domain expertise remains a challenge. In this paper, we address this challenge by proposing a systematic approach to learn execution contexts. We evaluate the effectiveness of the approach through extensive experiments on real-life event logs. An additional case study shows the usefulness of the approach to support workforce analytics in the context of process execution.","PeriodicalId":13255,"journal":{"name":"IEEE Transactions on Services Computing","volume":"19 5","pages":"3431-3446"},"PeriodicalIF":6.2,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148624010","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Performance and Security Aware Distributed Service Placement in Fog Computing","authors":"Mohammad Goudarzi;Arash Shaghaghi;Zhiyu Wang;Rajkumar Buyya","doi":"10.1109/TSC.2026.3725124","DOIUrl":"10.1109/TSC.2026.3725124","url":null,"abstract":"The rapid proliferation of Internet of Things (IoT) applications has intensified demand for efficient and secure service placement in Fog computing. However, heterogeneous resources, dynamic workloads, and diverse security requirements make optimal placement highly challenging, while most existing solutions prioritize performance metrics while overlooking security implications. This paper proposes a Security and Performance-Aware Distributed Deep Reinforcement Learning (SPA-DDRL) framework for joint optimization of service response time and security compliance. We formulate the problem as a weighted multi-objective optimization task, minimizing latency while maximizing a security score through a novel three-tier hierarchy: configuration-level checks verify proper settings, capability-level assessments evaluate resource security features, and control-level evaluations enforce stringent policies, strongly driving the policy toward compliant solutions aligned with performance objectives. SPA-DDRL adopts a distributed broker-learner architecture where multiple brokers perform autonomous placement decisions while a centralized learner coordinates global policy optimization through shared prioritized experiences. The framework integrates three key innovations: Long Short-Term Memory networks for temporal dependency modeling, Prioritized Experience Replay for accelerated convergence, and off-policy correction mechanisms for training stability. Extensive experiments demonstrate that SPA-DDRL achieves 16.3% response time improvement and 33% faster convergence compared with state-of-the-art approaches, while uniquely maintaining consistent, feasible security-compliant solutions across all system scales where baseline techniques fail or degrade.","PeriodicalId":13255,"journal":{"name":"IEEE Transactions on Services Computing","volume":"19 5","pages":"3554-3567"},"PeriodicalIF":6.2,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148755525","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"LUTless: Local Initial Approximation for Secure Power Function in Private Machine Learning","authors":"Xiangrui Xu;Qiao Zhang;Rui Ning;Chunsheng Xin;Hongyi Wu","doi":"10.1109/TSC.2026.3686664","DOIUrl":"10.1109/TSC.2026.3686664","url":null,"abstract":"The proliferation of deep learning service in privacy-sensitive domains demands frameworks that balance efficiency with model performance. However, modern secure multiparty computation (MPC) protocols struggle with prohibitive communication costs when evaluating non-linear operators essential to state-of-the-art architectures—specifically, Gaussian Error Linear Units (GeLU)/ReLU, Softmax, and LayerNorm in Transformers and CNNs. These functions often requires iterative power computations <inline-formula><tex-math>$f(x) = x^{1/a}$</tex-math></inline-formula>, where <inline-formula><tex-math>$a in mathcal {Z}_{ne 0,1}$</tex-math></inline-formula>, force a critical trade-off: precision-driven methods like Look-Up Tables (LUTs) require <inline-formula><tex-math>$O(2^lambda)$</tex-math></inline-formula> (<inline-formula><tex-math>$lambda$</tex-math></inline-formula> is bit-length) communication per approximation to preserve model accuracy, while efficiency-driven approaches using low-degree polynomials sacrifice accuracy on complex tasks due to unbounded approximation errors. To address these challenges, we propose LUTless, a novel method for secure power function that eliminates communication-intensive LUT operations while maintaining results accuracy by transferring communication load to local calculation during initial approximation. Our approach introduces double approximation technique that take advantage of the linear approximation of logarithm function and exponent approximation, successfully transfer heavy communication to local computation by reformulating power function. To meet the approximation requirements, we propose share flooding, a technique that pre-generates and securely masks shares to meet approximation requirements while preserving privacy. LUTless demonstrates broad applicability across machine learning tasks, including Transformer-based language models and CNNs. Evaluated on GLUE benchmark datasets for both encoder and decoder-based models, LUTless reduces communication by up to 3.9× and achieves 3.5× speedups while maintaining competitive accuracy compared to prior Transformer based inference works. For CNN inference, experiments with frameworks such as Cheetah and CrypTFlow2 demonstrate 2<inline-formula><tex-math>$times sim 6.9times$</tex-math></inline-formula> speedup and 1.6<inline-formula><tex-math>$times sim 9.7times$</tex-math></inline-formula> communication saving. Our results underscore the potential of LUTless to advance efficient and scalable privacy-preserving service across diverse applications.","PeriodicalId":13255,"journal":{"name":"IEEE Transactions on Services Computing","volume":"19 5","pages":"3144-3155"},"PeriodicalIF":6.2,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147735580","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Shielding Federated Learning Services From a Novel Stealthy Label-Flipping Attack Exposing Membership","authors":"Geetanjli Sharma;M.A.P. Chamikara;Mohan Baruwal Chhetri;Yi-Ping Phoebe Chen","doi":"10.1109/TSC.2026.3686135","DOIUrl":"10.1109/TSC.2026.3686135","url":null,"abstract":"Label-flipping attacks pose a significant threat to the security and privacy of machine learning models, particularly in federated learning (FL). While existing defenses, such as FoolsGold and Krum, aim to mitigate these attacks, they often fail to account for more subtle adversarial behaviors, such as intermittent class-targeted flips. In this work, we introduce the <bold>Stealthy Label-Flipping Attack (SLFA)</b>, which exposes this overlooked vulnerability in FL services. SLFA is an adaptive attack that intermittently flips a fixed subset of class labels across multiple communication rounds before reverting them to their original state. We show that this transient manipulation (i) evades state-of-the-art poisoning and anomaly detectors, (ii) preserves overall model utility, and (iii) introduces gradient instability during the attacking rounds, thereby increasing the risk of privacy leakage. To counter this hybrid threat that not only evades state-of-the-art defenses but also amplifies privacy leakage, we propose <bold>FLStealthShield</b>, a server-side defense mechanism that detects and neutralizes such adaptive attacks by analyzing temporal patterns in client updates. Extensive experiments under both IID and non-IID data distributions demonstrate that FLStealthShield achieves Byzantine robustness and maintains high model utility, outperforming existing defenses even when over 40% of clients are malicious.","PeriodicalId":13255,"journal":{"name":"IEEE Transactions on Services Computing","volume":"19 5","pages":"3308-3321"},"PeriodicalIF":6.2,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147731792","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Culexus: Goroutine-Granular Detection and Mitigation of Algorithmic Complexity Attacks in Go Microservices","authors":"Yihao Peng;Yiming Wu;Biao Ma;Risheng Wei;Hao Xi;Yue Chen;Hai Wan;Xibin Zhao","doi":"10.1109/TSC.2026.3706409","DOIUrl":"10.1109/TSC.2026.3706409","url":null,"abstract":"Go microservices leverage M:N goroutine-to-thread scheduling for high concurrency, but this model introduces critical vulnerabilities to Algorithmic Complexity Denial-of-Service (AC-DoS) attacks. Existing defenses target thread-per-request architectures; terminating an offending OS thread causes catastrophic collateral damage by killing thousands of co-located benign goroutines. We present Culexus, the first framework for goroutine-granular AC-DoS detection and mitigation in Go. Culexus injects request-specific identifiers into task contexts, then uses low-frequency kernel probes to statistically sample CPU activity, efficiently attributing resource consumption to individual requests. An offline training phase learns per-endpoint CPU cost thresholds from benign traffic; an in-kernel monitor then flags requests exceeding their learned budgets in real time. For mitigation, Culexus employs a novel two-tier adaptive strategy cooperating with the Go runtime: it first attempts surgical goroutine termination via the runtime’s native preemption mechanism, then automatically escalates for stubborn attacks (e.g., trapped in CGo calls) to a containment strategy that throttles the offending thread’s CPU while leveraging Go’s P Handoff mechanism to migrate benign requests to healthy threads. Evaluated on 16 diverse attack scenarios including 12 real-world CVEs, Culexus achieves 100% detection accuracy on our benchmark suite, millisecond-level surgical termination of preemptible attacks, limits benign traffic impact below 7% for stubborn attacks, and introduces less than 8% performance overhead.","PeriodicalId":13255,"journal":{"name":"IEEE Transactions on Services Computing","volume":"19 5","pages":"3416-3430"},"PeriodicalIF":6.2,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148314375","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}