近年来,入侵检测技术在网络安全中扮演着越来越重要的角色。目前的入侵检测模型所用的方法大部分是基于传统机器学习的浅层方法。浅层机器学习方法不能有效发掘数据特征,在入侵检测中存在一定的局限性。为此,论文提出了一种深度学习模型...近年来,入侵检测技术在网络安全中扮演着越来越重要的角色。目前的入侵检测模型所用的方法大部分是基于传统机器学习的浅层方法。浅层机器学习方法不能有效发掘数据特征,在入侵检测中存在一定的局限性。为此,论文提出了一种深度学习模型,该模型结合了多头注意力(multiHead attention)和双向门循环单元(BiGRU)。模型使用多头注意力和双向门循环单元从空间和时间上处理网络攻击流量,有效缓解模型复杂性,同时增加模型表现力。此外,使用最大池化方法(maxpooling)来平衡训练速度和性能,不但可以提取序列的边缘特征,还能帮助扩大感受野,由于数据不平衡会影响模型性能表现,因此使用随机过采样(Random Over Sampling)方法来处理数据不平衡的问题。实验基于UNSW-NB15数据集和CIC-IDS2017数据集,并使用准确率(Accuracy)、精确率(Precision)、召回率(Recall)和f1分数作为评估指标。实验结果表明,模型性能优秀。展开更多
With the rapid development of economy and transportation industry, highway construction is often faced with a lot of complex environment, if there is no reasonable economic management measures, then it is easy to redu...With the rapid development of economy and transportation industry, highway construction is often faced with a lot of complex environment, if there is no reasonable economic management measures, then it is easy to reduce the economic benefits of enterprises. Therefore, it is necessary for relevant staff to continuously improve the level of economic efficiency management, timely find the existing deficiencies, reduce the energy consumption of highway construction, achieve the improvement of highway economic efficiency, and enhance the competitiveness of enterprises.展开更多
文摘近年来,入侵检测技术在网络安全中扮演着越来越重要的角色。目前的入侵检测模型所用的方法大部分是基于传统机器学习的浅层方法。浅层机器学习方法不能有效发掘数据特征,在入侵检测中存在一定的局限性。为此,论文提出了一种深度学习模型,该模型结合了多头注意力(multiHead attention)和双向门循环单元(BiGRU)。模型使用多头注意力和双向门循环单元从空间和时间上处理网络攻击流量,有效缓解模型复杂性,同时增加模型表现力。此外,使用最大池化方法(maxpooling)来平衡训练速度和性能,不但可以提取序列的边缘特征,还能帮助扩大感受野,由于数据不平衡会影响模型性能表现,因此使用随机过采样(Random Over Sampling)方法来处理数据不平衡的问题。实验基于UNSW-NB15数据集和CIC-IDS2017数据集,并使用准确率(Accuracy)、精确率(Precision)、召回率(Recall)和f1分数作为评估指标。实验结果表明,模型性能优秀。
文摘With the rapid development of economy and transportation industry, highway construction is often faced with a lot of complex environment, if there is no reasonable economic management measures, then it is easy to reduce the economic benefits of enterprises. Therefore, it is necessary for relevant staff to continuously improve the level of economic efficiency management, timely find the existing deficiencies, reduce the energy consumption of highway construction, achieve the improvement of highway economic efficiency, and enhance the competitiveness of enterprises.
