COMPUTING PROJECTS

Topic: Malware/Ransomware/IoT Botnet Detection using Machine/Deep Learning
 

Description: We will investigate and implement an intelligent malware/ransomware/IoT botnet attack detection scheme. The signature‐based malware/ransomware/botnet detection methods will not be able to detect zero‐day and unknown malware/ransomware/botnet. Thus, a novel protection mechanism for malware/ransomware/botnet detection is needed and it should focus on malware/ransomware/botnet-specific operations to differentiate ransomware from other malware and benign files. This project will use a malware/ransomware/botnet detection method using an optimized version of deep learning through bio-inspired metaheuristics algorithms to achieve that purpose. Optimized versions of deep learning architectures like convolutional neural networks (CNNs) or other variants, can detect malware or ransomware or botnets efficiently simply by looking at the raw bytes of Windows/Linux Portable Executable files.
 

Skills needed: Programming skills in Java/C/C++/Python, Ability to read and understand research papers, basic AI/machine learning skills (need to pick up quickly) 

Topic: Sentiment Analysis of Twitter Feed using Machine Learning/Deep Learning

 

Description: There are many tweets surrounding an event. Many events have an important outcome like in the case of an election, mass protest or a football match. In this research, we will analyse the Twitter feeds surrounding an event using machine learning and will do sentiment analysis and/or behaviour prediction based on our analysis, along with other inputs surrounding that event. For example, we can consider tweets pertaining to specific teams and games in the football season and use them alongside statistical game data to build predictive models for future game outcomes or we could analyze the tweets of Donald Trump and see how he reacts to diverse topics.

 

Skills needed: Programming skills in Java/C/C++/MATLAB/Python, Ability to read and understand research papers, basic AI/machine learning skills (need to pick up quickly)

 

Topic: Digital Forensics Analysis using Machine Learning/Deep Learning

Network Forensics is the branch of Digital Forensics, where the evidence is network-related and exists in the form of logs, packets and network flows. Popular methods of investigating botnets include Honeypot, Network flow analysis, Intrusion detection systems, Visualization of Network traffic, Deep Packet Analysis etc. Multiple deep learning solutions have been proposed for application in the field of Network Forensics in recent years. Niyaz et al. (2016) used stacked auto-encoders in their implementation of a DDoS detection system for software-defined networks. The multiple auto-encoders were greedily trained layer-by-layer, with the output of one layer being the input of the next. Then the entire network was fine-tuned as a classifier. Lotfollahi et al. (2017) used a combination of a one-dimensional CNN and stacked auto-encoders for automatic feature extraction and classification of network traffic, achieving both application identification and traffic characterization in either encrypted or unencrypted traffic. This project will explore the use of Machine learning, Recurrent Neural Network (RNN), Convolutional Neural Networks (CNN), Deep Auto Encoder (DAE), Deep Boltzmann Machine (DBM) and Deep Belief Network (DBN), alongside some of the network forensics methods, whereby network attacks can be effectively mitigated.


Skills needed: Programming skills in Java/C/C++/MATLAB/Python, Ability to read and understand research papers, basic AI/machine learning skills (need to pick up quickly)

 

Topic: Intelligent Wireless Sensor Network Application in Healthcare

 

Description: Wireless inertial sensor network in conjunction with machine learning can be used in healthcare to work with elderly people to monitor their walk pattern and to alert a caregiver or relative when they have an unobtrusive fall. This will be more like an unobtrusive fall detection and alert system. The technology can also be used to monitor less critical patient's behaviour in a care home environment and can alert when an accident happens. The research will involve attaching wearable electronic sensors (depending on what is being monitored and which needs to be bought) on patients, collecting data wirelessly, analysing the data intelligently and taking action when anything abnormal occurs.

 

Skills needed: Network and wireless sensor network knowledge, Programming skills in Java/C/C++/MATLAB, Ability to read and understand research papers, basic machine learning skills (need to pick up quickly)

Energy-efficient Virtual Machine Placement using Bioinspired Optimisation Algorithms
 

The consolidation of virtual machines (VMs) helps to optimise the usage of resources and hence reduces the energy consumption in a cloud data centre. VM placement plays an important part in the consolidation of the VMs. The researchers have developed various algorithms for VM placement considering the optimised energy consumption. However, these algorithms lack the use of exploitation mechanisms efficiently. This project will look at VM placement issues by using existing and modified bioinspired meta-heuristic algorithms. The comparisons will be made against the existing algorithms, and the energy consumption results of all the participating algorithms will be looked at using CloudSim or similar simulation tools. The usage of the appropriate algorithm can help in efficient usage of energy in cloud computing.

 

Skills needed: Programming skills in Java/C/C++/MATLAB/Python, Ability to read and understand research papers, basic AI/machine learning/optimisation skills (need to pick up quickly)

Topic: Tackling Email Spamming through Machine Learning/Deep Learning

 

Description: Spam emails are causing major resource wastage by unnecessarily flooding the network links. Though many anti-spam solutions have been implemented, the Bayesian spam score approach looks quite promising, along with many other AI approaches. A proposal for spam detection algorithm can be done and its implementation using Java can be done with its performance test results on two independent spam corpora – Ling-spam and Enron-spam or various other spam datasets.

 

Skills needed: Programming skills in Java/C/C++/MATLAB/Python, Ability to read and understand research papers, basic AI/machine learning skills (need to pick up quickly)
 

Topic: Machine/Deep Learning in Network Security

 

Description: As network applications are increasingly being used with the popularity of broadband networks, network security is very important. The smartphones and computers use various applications for banking and online purchases, and the user needs to use it securely. To perform network attacks like spoofing, flooding, eavesdropping, etc. is so easy with some research. An intrusion detection system looks for different kind of network attacks in the incoming packets. Can we use machine learning to differentiate the attack packet flows, classify attacks and eventually stop them? In this work, we will try to use basic AI algorithms on different IDS datasets and find how effective they are. We will also look to improve their accuracy.

 

Skills needed: Programming skills in Java/C/C++/Python, Ability to read and understand research papers, basic AI/machine learning skills (need to pick up quickly)

 

Topic: Intelligent Analysis of Qualitative reports through Machine/Deep Learning

 

Description: To manually analyze a qualitative document can be time-consuming. For example, a corporate sustainability report is a report that is published by an organization about the economic, environmental and social impacts caused by its daily activities. These reports can help organizations to measure, understand and communicate the economic, environmental, social and governance performance. A manual analysis can be challenging when the report has hundred or more pages and when there are hundreds of such reports it can be even more challenging. In this research, we will use machine learning techniques along with natural language processing to do automated text analysis of (any) qualitative documents, in relation to their adherence to standard guidelines.
 

Skills needed: Programming skills in Java/C/C++/MATLAB/Python, Ability to read and understand research papers, basic AI/machine learning skills (need to pick up quickly)

Topic: Automatically Detect Eye Diseases (or Cancer Detection) through Machine/Deep Learning

 

Description: This project aims to automatically detect eye diseases early from retinal images through image processing and machine/deep learning. This work will use public datasets like DRIVE, STARE, and HRF. There is clearly a need to address and diagnose eye diseases early and efficiently which is the rationale behind this project. An estimated 19 million children under age 15, are visually impaired. Of these 12 million children are visually impaired due to refractive errors, a condition that could be easily diagnosed and corrected. NOTE: This project can be extended to cancer detection through processing CT Scan or MRI Scan images.

 

Skills needed: Programming skills in Java/C/C++/MATLAB/Python, Ability to read and understand research papers, basic machine learning skills (need to pick up quickly), basic image processing skills

 

Topic: Spam Detection in Online Social Networks Through Machine Learning/Deep Learning

 

Description: The main aim is to develop a machine learning based-detection system that can be used for different Online Social Network (OSN) platforms, such as Twitter and Facebook. Also, finding the best approach that can overcome spam drift issue in OSNs. The research objectives will be to implement the proposed approach through machine learning in Java/C++/Python to confirm the efficiency of the approach, to build large datasets with more features to be analyzed by the proposed approach, with the possible use of natural language processing and to improve the accuracy of the proposed approach through the use of optimization bio-inspired algorithms, such as ACO, Bee colony or firefly. Research Methodology can be as follows: (1) Data Collection: different OSN datasets will be built by using applications, such as Twitter API and Facebook API. (2) Feature Selection: Account-based, Content-based, Graph-based features will be extracted. (3) Preprocessing: Collected datasets are going to be cleaned, filtered, and split into different parts. (4) The use of natural language processing will be looked into, to find the exact word meanings within a context. (5) Experiments and Evaluations: Different experiments will be performed by using different algorithms, fine-tuning the parameters of those algorithms. Initially, WEKA is going to be used for evaluation. Then, the proposed approach is going to be evaluated by using Java or C++ to confirm the efficiency of the approach. Thus, a semi-supervised or other variants of ML approach will be used along with natural language processing to develop an optimized Machine Learning based system to detect spam across multiple OSNs.

 

Skills needed: Programming skills in Java/C/C++/Python, Ability to read and understand research papers, basic AI/machine learning skills (need to pick up quickly)

Topic: Penetration Testing and Security Attacks using Kali Linux

 

Description: The numbers of network hacking is increasing week after week and many companies are under attack. For this reason, many companies want security experts to do penetration testing or security assessment of their network. In this project we will try to use Kali Linux (penetration system tool) or other similar tools to attack other virtual machines and study the vulnerabilities in the system or in the protocols. The results and possible countermeasures will be studied and documented. The attacks can be extended to wireless networks as well.

 

Skills needed: Ability to work with virtual machines (servers or clients), do Linux scripting and learn/know how to use Kali Linux and tools in it.

Topic: Design of an Enterprise WAN Network

 

Description: WAN networks cover a wide geographical area. In this work, we will design a WAN which connects four locations in UK that constitutes four LANs. We will use Packet Tracer or real hardware to implement the network. We will implement different routing protocols like RIPv2, EIGRP and OSPF between routers, implement different VLANs within the LANs, configure Access Control Lists (standard and extended), configure a WAN protocol and configure DHCP and NAT. The PC in one network should be able to ping the PC in another network. We could study the performance of different routing protocols and see which one is efficient in our network. There are options for security attacks with ACLs configured.

 

Skills needed: Ability to work with Packet Tracer (and/or hardware routers and switches) and configure the network with advanced network skills