检索结果-内蒙古大学图书馆

On the Expressive Power of Behavior Structure

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Wang, Cheng Zhu, Hangyu Lin, Yuhang Jiang, Changjun Department of Computer Science and Technology Tongji University Shanghai201804 China Key Laboratory of Embedded System and Service Computing Ministry of Education Shanghai201804 China Shanghai Artificial Intelligence Laboratory Shanghai200232 China

Efforts toward a comprehensive description of behavior have indeed facilitated the development of representation-based approaches that utilize deep learning to capture behavioral information. As behavior complexity increases, the expressive power of these models reaches a bottleneck. We coin the term "behavioral molecular structure" and propose a new model called the Behavioral Molecular Structure (BMS). The model characterizes behaviors at the atomic level, analogizes behavioral attributes to atoms, and concretizes interrelations at the granularity of atoms using graphs. Here, we design three different downstream tasks to test the performance of the BMS model on public datasets. Additionally, we provide a preliminary theoretical analysis demonstrating that the BMS can offer effective expressiveness for complex behaviors. Copyright © 2023, The Authors. All rights reserved.

关键词： Atoms

Behavioral Authentication for Security and Safety

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Wang, Cheng Tang, Hao Zhu, Hangyu Zheng, Junhan Jiang, Changjun Department of Computer Science and Technology Tongji University Shanghai201804 China Key Laboratory of Embedded System and Service Computing Ministry of Education Shanghai201804 China Shanghai Artificial Intelligence Laboratory Shanghai200232 China

The issues of both system security and safety can be dissected integrally from the perspective of behavioral appropriateness. That is, a system is secure or safe can be judged by whether the behavior of certain agent(s) is appropriate or not. Specifically, a so-called appropriate behavior involves the right agent performing the right actions at the right time under certain conditions. Then, according to different levels of appropriateness and degrees of custodies, behavioral authentication can be graded into three levels, i.e., the authentication of behavioral Identity, Conformity, and Benignity. In a broad sense, for the security and safety issue, behavioral authentication is not only an innovative and promising method due to its inherent advantages but also a critical and fundamental problem due to the ubiquity of behavior generation and the necessity of behavior regulation in any system. By this classification, this review provides a comprehensive examination of the background and preliminaries of behavioral authentication. It further summarizes existing research based on their respective focus areas and characteristics. The challenges confronted by current behavioral authentication methods are analyzed, and potential research directions are discussed to promote the diversified and integrated development of behavioral authentication. Copyright © 2023, The Authors. All rights reserved.

关键词： Anomaly detection

Unsupervised Multivariate Time Series Anomaly Detection by Feature Decoupling in Federated Learning Scenarios

学校读者我要写书评

暂无评论

IEEE Transactions on Artificial Intelligence

IEEE Transactions on Artificial Intelligence 2025年

作者： He, Yifan Ding, Xi Tang, Yateng Guan, Jihong Zhou, Shuigeng Fudan University School of Computer Science Shanghai200433 China Shanghai Key Laboratory of Intelligent Information Processing Shanghai200433 China Tencent Weixin Group Guangzhou510308 China Tongji University Department of Computer Science and Technology Shanghai201804 China Ministry of Education Key Laboratory of Embedded System and Service Computing Tongji University Shanghai200092 China

Anomalies are usually regarded as data errors or novel patterns previously unseen, which are quite different from most observed data. Accurate detection of anomalies is crucial in various application scenarios. This paper focuses on unsupervised anomaly detection from Multivariate Time Series (MTS), as real-world data collected from sources such as wearable devices, medical equipment and industrial machines typically manifest as MTS and are often unlabeled. Anomaly detection in MTS represents a data-driven challenge that traditionally requires substantial centralized data for training models. However, in practice, data are frequently distributed among multiple institutions, with privacy concerns restricting unrestricted access. To address these issues, we introduce Feature Decoupling Federated Learning (FDFL), an approach designed to collaboratively train a representation learning network over multiple clients for unsupervised anomaly detection in MTS. Unlike previous methods that simply integrate MTS anomaly detection algorithms with federated learning strategies, FDFL specifically addresses heterogeneity among clients by decoupling the representation network into shared and private branches through a contrastive learning mechanism. This method aggregates shared parameters during each federated round while maintaining client-specific private parameters locally. Additionally, we develop a self-attention block that integrates the representations derived from both shared and private parameters to reconstruct MTS and identify anomalies based on reconstruction errors. Extensive experiments conducted on three publicly available datasets demonstrate that FDFL outperforms existing algorithms in most cases, highlighting the effectiveness and superiority of our proposed method in MTS anomaly detection. © 2020 IEEE.

关键词： Federated learning

Joint DNN Partition and Thread Allocation Optimization for Energy-Harvesting MEC systems

学校读者我要写书评

暂无评论

Joint DNN Partition and Thread Allocation Optimization for E...

Design, Automation and Test in Europe Conference and Exhibition

作者： Yizhou Shi Liying Li Yue Zeng Peijin Cong Junlong Zhou School of Computer Science and Engineering Nanjing University of Science and Technology Nanjing China National Key Laboratory for Novel Software Technology Nanjing University Nanjing China National Mobile Communications Research Laboratory Southeast University Nanjing China Key Laboratory of Embedded System and Service Computing (Tongji University) Ministry of Education Shanghai China

ISBN: (数字)9783982674100

ISBN: (纸本)9798331534646

Deep neural networks (DNNs) have demonstrated exceptional performance, leading to diverse applications across various mobile devices (MDs). Considering factors like portability and environmental sustainability, an increasing number of MDs are adopting energy harvesting (EH) techniques for power supply. However, the computational intensity of DNNs presents significant challenges for their deployment on these resource-constrained devices. Existing approaches often employ DNN partition or offloading to mitigate the time and energy consumption associated with running DNNs on MDs. Nonetheless, existing methods frequently fall short in accurately modeling the execution time of DNNs, and do not consider to use thread allocation for further latency and energy consumption optimization. To solve these problems, we propose a dynamic DNN partition and thread allocation method to optimize the latency and energy consumption of running DNNs on EH-enabled MDs. Specifically, we first investigate the relationship between DNN inference latency and allocated threads and establish an accurate DNN latency prediction model. Based on the prediction model, a DRL-based DNN partition (DDP) algorithm is designed to find the optimal partitions for DNNs. A thread allocation (TA) algorithm is proposed to reduce the inference latency. Experimental results from our test-bed platform demonstrate that compared to four benchmarking methods, our scheme can reduce DNN inference latency and energy consumption by up to 37.3% and 38.5%.

关键词： Energy consumption Instruction sets Heuristic algorithms Artificial neural networks Predictive models Prediction algorithms Inference algorithms Partitioning algorithms Resource management Optimization

Multi-modal Large Language Model Enhanced Pseudo 3D Perception Framework for Visual Commonsense Reasoning

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Zhu, Jian Wang, Hanli Shi, Miaojing The Department of Computer Science & Technology The Key Laboratory of Embedded System & Service Computing Ministry of Education Tongji University Shanghai200092 China The College of Electronic and Information Engineering Tongji University Shanghai201804 China

The visual commonsense reasoning (VCR) task is to choose an answer and provide a justifying rationale based on the given image and textural question. Representative works first recognize objects in images and then associate them with key words in texts. However, existing approaches do not consider exact positions of objects in a human-like three-dimensional (3D) manner, making them incompetent to accurately distinguish objects and understand visual relation. Recently, multi-modal large language models (MLLMs) have been used as powerful tools for several multi-modal tasks but not for VCR yet, which requires elaborate reasoning on specific visual objects referred by texts. In light of the above, an MLLM enhanced pseudo 3D perception framework is designed for VCR. Specifically, we first demonstrate that the relation between objects is relevant to object depths in images, and hence introduce object depth into VCR frameworks to infer 3D positions of objects in images. Then, a depth-aware Transformer is proposed to encode depth differences between objects into the attention mechanism of Transformer to discriminatively associate objects with visual scenes guided by depth. To further associate the answer with the depth of visual scene, each word in the answer is tagged with a pseudo depth to realize depth-aware association between answer words and objects. On the other hand, BLIP-2 as an MLLM is employed to process images and texts, and the referring expressions in texts involving specific visual objects are modified with linguistic object labels to serve as comprehensible MLLM inputs. Finally, a parameter optimization technique is devised to fully consider the quality of data batches based on multi-level reasoning confidence. Experiments on the VCR dataset demonstrate the superiority of the proposed framework over state-of-the-art approaches. Copyright © 2023, The Authors. All rights reserved.

关键词： Computational linguistics

LSR-IGRU: Stock Trend Prediction Based on Long Short-Term Relationships and Improved GRU

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Zhu, Peng Li, Yuante Hu, Yifan Liu, Qinyuan Cheng, Dawei Liang, Yuqi Department of Computer Science and Technology Tongji University Shanghai China Department of Computer Science and Technology Key Laboratory of Embedded System and Service Computing Ministry of Education Tongji University Artificial Intelligence Laboratory Shanghai China Seek Data Group Emoney Inc Shanghai China

Stock price prediction is a challenging problem in the field of finance and receives widespread attention. In recent years, with the rapid development of technologies such as deep learning and graph neural networks, more research methods have begun to focus on exploring the interrelationships between stocks. However, existing methods mostly focus on the short-term dynamic relationships of stocks and directly integrating relationship information with temporal information. They often overlook the complex nonlinear dynamic characteristics and potential higher-order interaction relationships among stocks in the stock market. Therefore, we propose a stock price trend prediction model named LSR-IGRU in this paper, which is based on long short-term stock relationships and an improved GRU input. Firstly, we construct a long short-term relationship matrix between stocks, where secondary industry information is employed for the first time to capture long-term relationships of stocks, and overnight price information is utilized to establish short-term relationships. Next, we improve the inputs of the GRU model at each step, enabling the model to more effectively integrate temporal information and long short-term relationship information, thereby significantly improving the accuracy of predicting stock trend changes. Finally, through extensive experiments on multiple datasets from stock markets in China and the United States, we validate the superiority of the proposed LSR-IGRU model over the current state-of-the-art baseline models. We also apply the proposed model to the algorithmic trading system of a financial company, achieving significantly higher cumulative portfolio returns compared to other baseline methods. Our sources are released at https://***/ZP1481616577/Baselines_LSR-IGRU. Copyright © 2024, The Authors. All rights reserved.

关键词： Graph neural networks

A Money Laundering Structure Detection Method Based on Incremental Transaction Analysis

学校读者我要写书评

暂无评论

A Money Laundering Structure Detection Method Based on Incre...

Data Science and Information Technology (DSIT), International Conference on

作者： Yifei Meng Zhong Li Key Laboratory of Embedded System and Service Computing Ministry of Education National (Province-Ministry Joint) Collaborative Innovation Center for Financial Network Security Tongji University Shanghai China College of Information Science and Technology Donghua University Shanghai China

Money laundering is the process of legitimizing dirty money through complex transactions, posing a serious threat to a country’s financial stability and national security. Nowadays, with the prevalence of organized money laundering, launderers prefer to use intricate multi-hop laundering chains to transfer dirty money. Moreover, they engage in normal financial activities to disrupt detection by auditors. In response to these trends and challenges, we propose a novel unsupervised money laundering structure detection framework for the anti-money laundering field, called structure incremental expansion (SIE). Our framework consists of two main modules: 1) Initialization of suspicious structures, which adopts a control-limit based method to identify suspicious accounts exhibiting anomalous transaction behavior. These accounts will serve as the starting points for suspicious structure expansion. 2) Dynamic structure expansion, where we design three dynamic membership functions according to the Financial Action Task Force’s definitions of the three stages of money laundering evolution. Newly-added incremental transactions in the network will be assigned to appropriate expanding suspicious structures. We conduct extensive experiments on simulated and public financial networks. SIE exhibits desirable performance and scalability. We also provide a detailed case study, visualizing a complete money laundering structure detection process, demonstrating our method’s strong interpretability.

关键词：

FedWNS: Data Distribution-Wise Node Selection in Federated Learning via Reinforcement Learning

学校读者我要写书评

暂无评论

FedWNS: Data Distribution-Wise Node Selection in Federated L...

International Conference on Computer supported Cooperative Work in Design

作者： Chengwu Tu Shengjie Zhao Hao Deng School of Software Engineering Tongji University Shanghai China Key Laboratory of Embedded System and Service Computing Shanghai China Engineering Research Center of Key Software Technologies for Smart City Perception and Planning Ministry of Education Shanghai China

To deal with the discrepancy between global and local objectives in the federated learning invoked by the non-independent, identically distributed (non-IID) data and mitigate the impact of catastrophic forgetting in the training phase, we propose a federated learning framework with data distribution-wise reinforcement learning to perform node selection to accelerate the convergence process and alleviate the accuracy degradation. In this framework, the agent on the central server observes the number of samples every node owns, the derived distribution information of every dataset, and the current local and global accuracy. Then infer the selected node-set to participate in the current federated learning round through policy network in reinforcement learning. Finally, we conduct simulations with publicly data sets. Simulation results indicate that our FedWNS outperforms the existing FedAvg and CSFedAvg on the testing accuracy and the communication rounds to reach target accuracy under different settings.

关键词：

MIME: Fast and Accurate Flow Information Compression for Multi-Spread Estimation

学校读者我要写书评

暂无评论

MIME: Fast and Accurate Flow Information Compression for Mul...

International Conference on Network Protocols

作者： Hanwen Zhang He Huang Yu-E Sun Zhaojie Wang School of Computer Science & Technology Soochow University Suzhou China School of Rail Transportation Soochow University Suzhou China Key Laboratory of Embedded System and Service Computing Ministry of Education Tongji University Shanghai China

Spread estimation is an essential issue in high-speed networks with wide applications, such as network billing, quality of service, anomaly detection, etc. As a promising technique, sketch can efficiently estimate per-flow spread with only a small memory cost. Many studies focus on improving the performance of sketches. However, these works primarily focus on optimizing the counter level or sketch level without considering the scenario of multi-spread estimation, which is crucial for improving memory utilization and detecting anomalies. In this paper, we propose an efficient flow information compression algorithm based on the on-chip/off-chip hybrid framework to estimate multiple flow spreads. In the on-chip memory, we filter out non-duplicates and sample them to the off-chip space for recording. In the off-chip memory, we compress each sampled non-duplicate to a carefully designed bit-cube. When the measurement is finished, we separate corresponding KV-flows from a specific flow based on the user query. Then, we rebuild this flow to a subset group based on the duplicate number of each KV-flow. Finally, according to the Multi-set theory, we derive an accurate multi-spread estimate formula to solve this issue with a high throughput and small on-chip memory usage. Furthermore, we evaluate the performance of our proposed estimator using real Internet traffic traces downloaded from CAIDA. Experiments show that, compared to the state-of-the-art, our proposal achieves a 97.2% lower average relative error in per-destination source flow spread estimation with a tight on-chip memory, e.g., 320KB. And our proposed method achieves 31.86 higher processing throughput.

关键词：