检索结果-内蒙古大学图书馆

Downstream-agnostic Adversarial Examples

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Zhou, Ziqi Hu, Shengshan Zhao, Ruizhi Wang, Qian Zhang, Leo Yu Hou, Junhui Jin, Hai School of Cyber Science and Engineering Huazhong University of Science and Technology China School of Computer Science and Technology Huazhong University of Science and Technology China School of Cyber Science and Engineering Wuhan University China School of Information and Communication Technology Griffith University Australia Department of Computer Science City University of Hong Kong Hong Kong National Engineering Research Center for Big Data Technology and System Services Computing Technology and System Lab Hubei Key Laboratory of Distributed System Security China Hubei Engineering Research Center on Big Data Security China Cluster and Grid Computing Lab

Self-supervised learning usually uses a large amount of unlabeled data to pre-train an encoder which can be used as a general-purpose feature extractor, such that downstream users only need to perform fine-tuning operations to enjoy the benefit of "large model". Despite this promising prospect, the security of pre-trained encoder has not been thoroughly investigated yet, especially when the pre-trained encoder is publicly available for commercial use. In this paper, we propose AdvEncoder, the first framework for generating downstream-agnostic universal adversarial examples based on the pre-trained encoder. AdvEncoder aims to construct a universal adversarial perturbation or patch for a set of natural images that can fool all the downstream tasks inheriting the victim pre-trained encoder. Unlike traditional adversarial example works, the pre-trained encoder only outputs feature vectors rather than classification labels. Therefore, we first exploit the high frequency component information of the image to guide the generation of adversarial examples. Then we design a generative attack framework to construct adversarial perturbations/patches by learning the distribution of the attack surrogate dataset to improve their attack success rates and transferability. Our results show that an attacker can successfully attack downstream tasks without knowing either the pre-training dataset or the downstream dataset. We also tailor four defenses for pre-trained encoders, the results of which further prove the attack ability of AdvEncoder. Our codes are available at: https://github. com/CGCL-codes/AdvEncoder. Copyright © 2023, The Authors. All rights reserved.

关键词： Machine learning

Encoding Global Semantic and Localized Geographic Spatial-Temporal Relations for Traffic Accident Risk Prediction

学校读者我要写书评

暂无评论

SSRN

SSRN 2024年

作者： Alhaek, Fares Li, Tianrui Rajeh, Taha M. Javed, Muhammad Hafeez Liang, Weichao School of Computing and Artificial Intelligence Southwest Jiaotong University Chengdu611756 China Engineering Research Center of Sustainable Urban Intelligent Transportation Ministry of Education Chengdu611756 China National Engineering Laboratory of Integrated Transportation Big Data Application Technology Southwest Jiaotong University Chengdu611756 China Manufacturing Industry Chains Collaboration and Information Support Technology Key Laboratory of Sichuan Province Southwest Jiaotong University Chengdu611756 China

The proliferation of vehicles and the intricate layout of road systems have contributed to a significant rise in traffic accidents, posing a pressing concern globally. Despite the advancements facilitated by deep learning, several challenges persist in the domain of traffic accident prediction. Firstly, the sparsity of accident data in certain city regions presents a significant obstacle, particularly when attempting fine-grained predictions at a local level. Secondly, the intricate spatial-temporal relations inherent in traffic accident data pose a challenge for existing prediction models. In response to these formidable challenges, this paper presents GLST-TARP, a novel model for predicting traffic accident risk in urban environments by leveraging both global semantic and localized geographic spatial-temporal relations. Specifically, we construct multi-graphs to encode global static and dynamic spatial relations, incorporating attention mechanisms to adaptively focus on relevant information and temporal relations. Additionally, we employ channel-wise convolutional neural network blocks to extract localized geographic features and enhance predictive accuracy. The proposed model is trained using a Huber loss function tailored for regression tasks, relieving the impact of zero values during optimization. Experimental results demonstrate that GLST-TARP outperforms state-of-the-art methods in predicting traffic accident risks, showcasing its potential for enhancing urban safety and transportation management. © 2024, The Authors. All rights reserved.

关键词： Graph neural networks

Temporal Knowledge Graph Reasoning via Time-Distributed Representation Learning

学校读者我要写书评

暂无评论

Temporal Knowledge Graph Reasoning via Time-Distributed Repr...

IEEE International Conference on data Mining (ICDM)

作者： Kangzheng Liu Feng Zhao Guandong Xu Xianzhi Wang Hai Jin National Engineering Research Center for Big Data Technology and System Services Computing Technology and System Lab Cluster and Grid Computing Lab School of Computer Science and Technology Huazhong University of Science and Technology Wuhan China Data Science and Machine Intelligence Lab University of Technology Sydney Sydney Australia

Temporal knowledge graph (TKG) reasoning has attracted significant attention. Recent approaches for modeling historical information have led to great advances. However, the problems of time variability and unseen entities have become two major obstacles preventing further development. The time variability problem means that different historical timestamps play different roles in the inference process. Furthermore, in the context of time variability, the unseen entity problem means that a query cannot obtain a predicted entity that is unseen in the scale-varying history rather than in a fixed set, thus turning from static to dynamic. In this paper, we propose a novel method named DHU-NET for addressing the time variability challenge and the dynamic unseen entity challenge derived from it. With regard to the former concern, we propose a time-distributed representation learning method based on a graph convolutional network(GCN) and a self-attention mechanism, which learns the distributed representations of facts at different historical timestamps and comprehensively pays different levels of attention to the different timestamps. With regard to the latter issue, we extract the unseen entities from a global static KG based on a copy mechanism and bring them into consideration during the final prediction step. Experiments on six benchmark datasets demonstrate the substantial improvements achieved by DHUNET in terms of multiple evaluation metrics. Our released codes are available at https://***/CGCL-codes/DHUNET.

关键词： Representation learning Measurement Knowledge engineering Predictive models Turning Cognition data models

GIDN: A Lightweight Graph Inception Diffusion Network for High-efficient Link Prediction

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Wang, Zixiao Guo, Yuluo Zhao, Jin Zhang, Yu Yu, Hui Liao, Xiaofei Wang, Biao Yu, Ting National Engineering Research Center for Big Data Technology and System Services Computing Technology and System Lab Cluster and Grid Computing Lab School of Computer Science and Technology Huazhong University of Science and Technology Wuhan China Zhejiang Lab Hangzhou China

In this paper, we propose a Graph Inception Diffusion Networks(GIDN) model. This model generalizes graph diffusion in different feature spaces, and uses the inception module to avoid the large amount of computations caused by complex network structures. We evaluate GIDN model on Open Graph Benchmark(OGB) datasets, reached an 11% higher performance than AGDN on ogbl-collab dataset. Copyright © 2022, The Authors. All rights reserved.

关键词： Complex networks

PROM: A persistent routing optimization method based on supervised learning

学校读者我要写书评

暂无评论

Journal of Network and Computer Applications 2025年 242卷

作者： Yingya Guo Zebo Huang Mingjie Ding Bin Lin Huan Luo College of Computer and Data Science Fuzhou University China Fujian Provincial Key Laboratory of Network Computing and Intelligent Information Processing Fuzhou University China Engineering Research Center of Big Data Intelligence Ministry of Education China

Traditional Traffic Engineering (TE) typically improves network performance by formulating and solving routing optimization problems based on measured or predicted Traffic Matrices (TMs). However, since network traffic is dynamic, routing strategies optimized for the current measured TMs may not be suitable for future traffic patterns. To adapt to dynamic traffic, one possible solution is to perform more frequent routing updates, but this may lead to service disruptions. Another approach is the two-step method of prediction followed by optimization, which could lead to degraded performance of routing strategies due to inaccuracies in traffic prediction. Additionally, existing approaches focus solely on either distributed or centralized networks, lacking flexibility to adapt to different network architectures. To address these challenges, we propose PROM, a Machine Learning (ML)-based routing optimization model that directly predicts an effective routing strategy in an end-to-end manner based on historical traffic. This routing strategy is applicable across multiple future TMs, thereby reducing the need for frequent routing updates. Furthermore, PROM can be easily extended from a centralized implementation to a distributed one, adapting to different network architectures. To enhance model performance, we have also designed a custom loss function for routing optimization scenarios to avoid overfitting. Extensive simulation results demonstrate that PROM can predict high-quality routing strategies in future unknown traffic scenarios and is adaptable to both centralized and distributed network architectures.

关键词：

Which Classifier to Deploy? A Hybrid Sampling Approach for Evaluating Classifiers with Unbalanced data in Medical Expert Systems

学校读者我要写书评

暂无评论

SSRN

SSRN 2023年

作者： Yu, Haiyan Qiu, Hang Piri, Saeed Miao, Hongxia Center for Data and Decision Sciences School of Economics and Management Chongqing University of Posts and Telecommunications Chongqing400065 China Health Big Data Research Institute School of Computing and Engineering University of Electronic Science and Technology of China Chengdu611731 China Department of Operations and Business Analytics Lundquist College of Business University of Oregon EugeneOR97403 United States

To develop medical expert systems (MES), researchers and practitioners usually apply a machine learning (ML) classifier they expect to be the best through ML effect estimation. A standard classifier can perform well on balanced datasets, but its performance often dramatically decreases when applied to unbalanced datasets. To address these challenges, this paper proposes a hybrid sampling approach for evaluating classifiers with unbalanced data in MES. In this framework, the mean responses of the classification experiments are employed to estimate the ML effect. Meanwhile, uniform designs are used in hybrid sampling (HSUD) to retrieve the optimal parameters. The efficacy of HSUD is validated through applications on six real-world clinical datasets (unbalanced) using six ML algorithms (including zero rule). The ratio of the minority to the majority ranges from 0.09 to 0.80 in the six datasets, and the sample size ranges from 187 to 15151. The results suggest that our approach is more effective than existing unbalanced data learning methods. In particular, the AUC with HSUD is 2.48% and 8.60% higher than those with proportional and balanced sampling methods in the testing, respectively. Moreover, HSUD is efficient in choosing classifiers with unbalanced data in MES. © 2023, The Authors. All rights reserved.

关键词： Machine learning

Iterative Refinement of Project-Level Code Context for Precise Code Generation with Compiler Feedback

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Bi, Zhangqian Wan, Yao Wang, Zheng Zhang, Hongyu Guan, Batu Lu, Fangxin Zhang, Zili Sui, Yulei Jin, Hai Shi, Xuanhua Huazhong University of Science and Technology China University of Leeds United Kingdom Chongqing University China Shanghai Jiao Tong University China University of New South Wales Australia National Engineering Research Center for Big Data Technology and System Services Computing Technology and System Lab Cluster and Grid Computing Lab School of Computer Science and Technology Huazhong University of Science and Technology Wuhan430074 China

Large Language Models (LLMs) have shown remarkable progress in automated code generation. Yet, LLM-generated code may contain errors in API usage, class, data structure, or missing project-specific information. As much of this project-specific context cannot fit into the prompts of LLMs, we must find ways to allow the model to explore the project-level code context. We present COCOGEN, a new code generation approach that uses compiler feedback to improve the LLM-generated code. COCOGEN first leverages static analysis to identify mismatches between the generated code and the project’s context. It then iteratively aligns and fixes the identified errors using information extracted from the code repository. We integrate COCOGEN with two representative LLMs, i.e., GPT-3.5-Turbo and Code Llama (13B), and apply it to Python code generation. Experimental results show that COCOGEN significantly improves the vanilla LLMs by over 80% in generating code dependent on the project context and consistently outperforms the existing retrieval-based code generation baselines. © 2024, CC BY-NC-SA.

关键词： Static analysis

The importance of robust and reliable energy prediction models: Next generation of smart meters 10

学校读者我要写书评

暂无评论

The importance of robust and reliable energy prediction mode...

10th International Conference on Simulation and Modeling Methodologies, Technologies and Applications, SIMULTECH 2020

作者： Jurado, Sergio Nebot, Àngela Mugica, Francisco Soft Computing Research Group Intelligent Data Science and Artificial Intelligence Research Center Universitat Politècnica de Catalunya Barcelona Spain

ISBN: (纸本)9789897584442

In this position paper a discussion is performed related to the importance of Energy Prediction Models (EPM) in the context of trading renewable energy in smart grids and the need to develop a new generation of Smart Meters (SMs) based on edge computing. If the electricity currently produced and consumed in the low voltage is expected to grow due to installations of PVs and the electrification of the system, we need to incentivise local energy trading and provide tools to make it possible. Currently, all energy trading mechanisms in the literature heavily rely on predictions, therefore, inaccuracy would be translated in low profitability of prosumer's investments or even worse, disappointment with the local energy markets. To guarantee robustness and reliability of predictions, we propose Flexible FIR as EPM to be used during the trading process and to integrate them in a Next Generation of SMs (NGSMs). In this document some reflections about new functionalities of NGSMs and first steps of a prototype are also addressed. © 2020 by SCITEPRESS - Science and Technology Publications, Lda. All rights reserved

关键词： Smart meters

a branch-and-price approach for the nurse rostering problem with multiple units

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Hu, Wanzhe He, Xiaozhou Luo, Li Pardalos, Panos M. School of Economics and Management Chongqing University of Posts and Telecommunications Chongqing China Business School Sichuan University Sichuan Chengdu China Key Laboratory of Big Data Intelligent Computing Chongqing University of Posts and Telecommunications Chongqing China Department of Industrial and Systems Engineering Center for Applied Optimization University of Florida Gainesville United States

In this paper, we study the nurse rostering problem that considers multiple units and many soft time-related constraints. An efficient branch and price solution approach that relies on a fast algorithm to solve the pricing subproblem of the column generation process is presented. For the nurse rostering problem, its pricing subproblem can be formulated as a shortest path problem with resource constraints, which has been the backbone of several solutions for several classical problems like vehicle routing problems. However, approaches that perform well on these problems cannot be used since most constraints in the nurse rostering problem are soft. Based on ideas borrowed from global constraints in constraint programming to model rostering problems, an efficient dynamic programming algorithm with novel label definitions and dominating rules specific to soft time-related constraints is proposed. In addition, several acceleration strategies are employed to improve the branch and price algorithm. Computational results on instances of different sizes indicate that the proposed algorithm is a promising solution for the nurse rostering problem with multiple units. Copyright © 2023, The Authors. All rights reserved.

关键词： Constraint programming