Scribd Logo
Upload

Welcome to Scribd!

  • 45 views

    Presented by Name: Roll No:: Bukke Pratap Kumar Naik 17121A0520

    Uploaded by

    Dronadula Venkata Sumanth Sumanth
    This document presents a seminar on cyber protection by deep learning in IoT. It discusses using deep learning approaches like TensorFlow neural networks to detect software piracy and malware threats in IoT networks. It proposes an architecture model and applies techniques like TFIDF weighting, convolutional neural networks, and color image visualization on malware files to detect threats. The results show accuracy rates of over 97% for piracy detection and malware classification. The conclusion discusses challenges of security in growing IoT networks and how the combined deep learning approach can help identify pirated and malware files.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd

    Presented by Name: Roll No:: Bukke Pratap Kumar Naik 17121A0520

    Uploaded by

    Dronadula Venkata Sumanth Sumanth
    45 views15 pages
    This document presents a seminar on cyber protection by deep learning in IoT. It discusses using deep learning approaches like TensorFlow neural networks to detect software piracy and malware threats in IoT networks. It proposes an architecture model and applies techniques like TFIDF weighting, convolutional neural networks, and color image visualization on malware files to detect threats. The results show accuracy rates of over 97% for piracy detection and malware classification. The conclusion discusses challenges of security in growing IoT networks and how the combined deep learning approach can help identify pirated and malware files.

    Original Description:

    Guuhbb

    Original Title

    17121A0520

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    This document presents a seminar on cyber protection by deep learning in IoT. It discusses using deep learning approaches like TensorFlow neural networks to detect software piracy and malware threats in IoT networks. It proposes an architecture model and applies techniques like TFIDF weighting, convolutional neural networks, and color image visualization on malware files to detect threats. The results show accuracy rates of over 97% for piracy detection and malware classification. The conclusion discusses challenges of security in growing IoT networks and how the combined deep learning approach can help identify pirated and malware files.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    45 views15 pages

    Presented by Name: Roll No:: Bukke Pratap Kumar Naik 17121A0520

    Uploaded by

    Dronadula Venkata Sumanth Sumanth
    This document presents a seminar on cyber protection by deep learning in IoT. It discusses using deep learning approaches like TensorFlow neural networks to detect software piracy and malware threats in IoT networks. It proposes an architecture model and applies techniques like TFIDF weighting, convolutional neural networks, and color image visualization on malware files to detect threats. The results show accuracy rates of over 97% for piracy detection and malware classification. The conclusion discusses challenges of security in growing IoT networks and how the combined deep learning approach can help identify pirated and malware files.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    of 15

    A SEMINAR ON

    Cyber Protection by DeepLearning in IOT

    Presented by

    Name: BUKKE PRATAP KUMAR NAIK


    Roll No:17121A0520

    Under the guidance of HOD:


    Dr.G.Sunitha Dr. B.Narendra Kumar Rao,
    (Associate Professor) Professor & Head,
    Dept. of CSE

    Computer Science and Engineering


    CONTENTS
    Ø Abstract
    Ø Introduction
    Ø Objectives
    Ø Scope
    Ø Literature study
    Ø Algorithm or technique
    Ø Observed results
    Ø Conclusion
    Ø References
    ABSTRACT
    • The IoT (Internet of Things) connect systems, applications, data storage,
    and services that may be a new gateway for cyber-attacks as they
    constantly offer services in the organization.
    • Currently, software piracy and malware attacks are high risks to
    compromise the security of IoT. These threats may steal important
    information that causes economic and reputational damages
    • In this paper, we have proposed a combined deep learning approach to
    detect the pirated software and malware infected files across the IoT
    network
    • The TensorFlow deep neural network is proposed to identify pirated
    software using source code plagiarism
    • the deep learning approach is used to detect source code plagiarism. The
    dataset is collected from Google Code Jam (GCJ) to investigate software
    piracy
    INTRODUCTION

    • The Internet of Things (IoT) is the interconnection of physical


    moving objects “Things” through internet embedded with an
    electronic chip, sensors, and other forms of hardware
    • Each device is uniquely identified globally by Radio Frequency
    Identifier (RFID) tags.
    • CHALLENGES
    • Software Piracy:
    – Currently, every third installed software application is pirated
    – The intellectual digital property and authorship rights of software are not
    physical due to globally accessible on the internet and thus, hard to sustain
    – The attacker may crack the original software and re-design the logic into
    another type of programming language.
    INTRODUCTION

    • Malware detection:
    – The conventional methods may solve code obfuscation
    concerns, but high computational cost is needed
    – How to identify malware with reduced overhead?
    – How to extract malware features with less computational
    cost?
    – How to process big malware datasets to get better
    accuracy?
    • In this paper, we proposed a combined deep learning
    approach to identify the pirated and Malware attacks on
    industrial IoT cloud
    OBJECTIVES

    • High-speed querying of security intelligence data.


    • Flexible big data analytics across structured as well as
    unstructured data
    • Graphical front-end tool for visualizing as well as exploring big
    data.
    • Integrity ,authenticity,availabity access and confidentialty
    • Risk management and analysis
    • Data protection
    • Helps to fight against cyber-crimes towards predicting and in
    preventing them.
    SCOPE

    There is a scope for Bigdata with Enhancing Cyber Security


    • Health care
    • Product improvement
    • Environment study
    • Education improvements
    • It sector
    • Social networking apps and websites
    • Reducing terrorist attacks
    • Providing security by means of authentication
    • Gadget like phones,pc..etc are to be protected which are
    connected to internet
    LITERATURE STUDY

    • For efficient threat detection in the IoT environment, more studies have
    been conducted on malware detection and software plagiarism detection to
    achieve high Identification performance, and reduce time cost.
    • Several studies have observed the impact of different aspects of software
    plagiarism detection
    • Most of the previous work is complete on a single programming language
    • , the author proposed a Source Forager search engine that was used to
    extract like functionalities between C and C++code in response to user
    questions
    • the logic-based approach was used to compute the behavior of dissimilarity
    in two source codes by the same inputs. If there was no dissimilarity, then it
    originated in plagiarism problem
    ALGORITHM OR TECHNIQUE

    architecture model for cyber security threats and protection measures in


    industrial IoT
    MALWARE THREAT DETECTION
    • Data preprocessing
    – The color images are generated from raw binary files to transform the malware
    detection problem using image classification problem
    • DEEP CONVOLUTIONAL NEURAL NETWORK
    – The Deep Convolutional Neural Network (DCNN) is proposed to conduct in-depth
    malware data analysis.
    – The input layer is used to receive training images for the designed neural network
    model.
    • CONVOLUTION LAYER
    – The meaningful features are extracted by reducing the image of parameters using
    convolution layer
    – . It decreases the over-fitting problem and gives the generalization concept to the main
    architecture
    ALGORITHM OR TECHNIQUE

    • POOLING LAYER
    – The sub-sampling layer term is generally used for pooling layer which offers
    two modes of pooling, i.e., maximum and avg pooling
    – to minimize the loss incurred on the training data.
    SOFTWARE PIRACY THREAT DETECTION
    • PRE-PROCESSING AND FEATURE EXTRACTION
    • DEEP LEARNING WITH TENSORFLOW FRAMEWORK
    – The TensorFlow is a machine learning system which is used for high-level computations
    in a complex environment.
    – We can implement different types of machine and deep learning algorithms by calling
    specific Application Programing Interface (API) of TensorFlow.
    – It has different types of layers which can be configured for complex computations,
    training the data, and supervise the state-run of each function
    ALGORITHM OR TECHNIQUE

    Architecture model for Cyber Security Threats prediction in IoT


    OBSERVED RESULTS

    • The software plagiarism measure may be used to investigate the code


    similarity in pirated software
    • We took source code dataset from Google Code Jam (GCJ) to analyze the
    proposed methodology for software piracy.
    • Term Frequency and Inverse Document Frequency (TFIDF) and Logarithm
    Term Frequency (LogTF) are used to retrieve the tokens’ weighting.
    • The box plot of weighting values is shown in figure 2. The P1, P2, P3, and
    P4 on x-axis denote the programming solutions of four solved questions,
    respectively
    • , GIST, LBP, and CLGM descriptors with SVM classifier.
    • accuracy obtained by the proposed approach was 97.46%,
    • GIST + SVM was 86.1%, LBP + SVM
    • was 78.05%, and CLGM + SVM was 92.06% respectively
    OBSERVED RESULTS

    Box plot of source codes weighting values

    The highest F-measure obtained by the proposed method was 97.44%, GIST + SVM was
    85.82%, LBP + SVM was 77.49%, and CLGM + SVM was
    91.98% respectively.
    CONCLUSION

    • The industrial IoT based network is rapidly growing in the coming future
    • The detection of software piracy and malware threats are the main
    challenges in the field of cybersecurity using IoT-based big data
    • We proposed a combined deep learning based approach for the
    identification of pirated and malware file
    • . The source code is preprocessed to clean from noise and to capture
    further the high-quality features which include useful tokens
    • Then, TFIDF and LogTF weighting techniques are used to zoom the
    contribution of each token in terms of source code similarity
    • a novel methodology based on convolution neural network and color
    image visualization to detect malware using IoT.
    REFERENCES

    • Srinivasan, C., et al. A review on the different types of internet of


    things (IoT). Journal of Advanced Research in Dynamical and
    Control Systems, 2019. 11(1): p. 154-158.
    • Joyia, G.J., et al., Internet of Medical Things (IOMT): applications,
    benefits and future challenges in healthcare domain. J Commun,
    2017. 12(4): p. 240-7.
    • Zanella, A., et al., Internet of things for smart cities. IEEE Internet of
    Things journal, 2014. 1(1): p. 22-32.
    • Karbab, E.B., et al., Android malware detection using deep learning
    on api method sequences. arXiv preprint arXiv:1712.08996, 2017.
    • Jabbar, S., et al., A methodology of real-time data fusion for
    localized big data analytics. IEEE Access, 2018. 6: p. 24510- 24520.

    You might also like