دانلود رایگان مقاله RNNIDS: تقویت سیستم های تشخیص نفوذ شبکه از طریق یادگیری عمیق – سال 2021
مشخصات مقاله:
عنوان فارسی مقاله:
RNNIDS: تقویت سیستم های تشخیص نفوذ شبکه از طریق یادگیری عمیق
عنوان انگلیسی مقاله:
RNNIDS: Enhancing Network Intrusion Detection Systems through Deep Learning
سال انتشار مقاله:
2021
کلمات کلیدی مقاله:
امنیت شبکه، سیستم های تشخیص نفوذ شبکه، کرم، جهش ها تولید Dataset، یادگیری عمیق، شبکه های عصبی راجعه
مناسب برای رشته های دانشگاهی زیر:
کامپیوتر
مناسب برای گرایش های دانشگاهی زیر:
مهندسی نرم افزار، امنیت اطلاعات
وضعیت مقاله انگلیسی و ترجمه:
مقاله انگلیسی را میتوانید به صورت رایگان با فرمت PDF با کلیک بر روی دکمه آبی، دانلود نمایید. برای ثبت سفارش ترجمه نیز روی دکلمه قرمز رنگ کلیک نمایید. سفارش ترجمه نیازمند زمان بوده و ترجمه این مقاله آماده نمیباشد و پس از اتمام ترجمه، فایل ورد تایپ شده قابل دانلود خواهد بود.
فهرست مطالب:
Abstract
Keywords
1. Introduction
2. Notations and definitions
3. Brief overview of recurrent neural networks (RNNs)
3.1. Application of RNNs in text generation
4. Methodology: applications of RNNs in intrusion detection
4.1. Generating new mutants of polymorphic worm
4.2. Generating synthetic signatures
5. Experiment description
5.1. Metrics
5.2. Setup of the experiments
5.3. Experiment design
6. Results and discussion
6.1. Similarity between original and RNN-generated Worms
6.2. Similarity between original and RNN-generated Bro Signatures
6.3. Enhancing the performance of Bro
7. Conclusion and remarks
Declaration of Competing Interest
Acknowledgements
References
قسمتی از مقاله انگلیسی:
1 INTRODUCTION
Nowadays we are witnessing rapidly escalating Internet threats, which have become increasingly mature as the Internet and its applications evolve. Today’s Internet provides ubiquitous connectivity to a wide range of devices, with different operating systems, which indeed expands the available attack surface including several different attack vectors. As a prime example, according to the recent Symantec report1 , a significant increase can be observed in different classes of attacks, e.g., internet of things (IoT) devices (more than 600%), new downloader variants (more than 92%), etc. [58]. This increase has been partially fueled by the increased availability of user-friendly hacking tools, demanding solely superficial knowledge from attackers, as illustrated first by Lipson [39] and further extended in [11]. Among malicious activities practiced by attackers, several classes of attacks can be recognized, for instance, Denial of Service (DoS) [40], disclosure, manipulation impersonation, and repudiation. These classes can be lumped together by an umbrella term, namely intrusion. Along with the emergence of increasingly sophisticated intrusions, intrusion detection systems (IDS) have been developed to cope with these threats. Regarding where or at which point an IDS is placed, two types of such systems can be distinguished: network intrusion detection systems (NIDS), and host intrusion detection systems (HIDS) [14]. The latter is run on a device or an individual host in the network, whereas an NIDS is located within the network, at a strategic point, to monitor the traffic to and from all devices. Irrespective of this classification, the intrusion detection systems share some commonalities; first, they take advantage of the connectivity provided by networks, and secondly, they either use the known, specific patterns or apply anomaly detection techniques.
