COMP40511 Applied Artificial Intelligence Coursework Assignment Brief | NTU

Module Learning outcomes COMP40511 

MLO1 Demonstrate an in-depth understanding of the theory behind AI and data mining techniques and their critical evaluation.
MLO2 Critically appraise the main issues involved in real-world AI and data mining systems and the consequences of their use.
MLO3 Design, implement and critically evaluate machine learning systems to solve real-world problems.
MLO4 Critically appraise and apply AI technology to extract useful and actionable insight from data and use it to assist decision- making.
MLO5 Write a detailed technical report describing the solutions.

I. Assessment Overview

The aim of this assignment is to demonstrate knowledge and understanding of the applications of artificial intelligence techniques in different scenarios/ case studies, according to the specifications given below.

You are required to produce a complete report and python source code for your development using the specification given below.

Plagiarism Declaration

The following PLAGIARISM DECLARATION must be entered into the report and the name and ID completed:
“This report and the source code it documents are the result of my own work. Any contributions to the work by third parties, other than tutors, are stated clearly below this declaration. Should this statement prove to be untrue I recognise the right and duty of the Board of Examiners to take appropriate action in line with the university’s regulations on assessment.
Name:………………………………………………………………………………………….ID No    ”

Good luck and have fun!

II.  Assessment Requirements

Each student must choose a topic/problem in a subject area of interest that they can obtain a reasonable dataset to be used in implementing a few solutions using the prescribed AI techniques namely, Artificial Neural Networks, Deep CNNs, Support Vector Machines and Clustering approaches. The chosen topic/problem to be addressed should be a real-world problem that can benefit from an AI-based solution. You can discuss your topic with your tutor during lab sessions.

Each student is required to produce a solution for the following assessment components:

1. Project Proposal (10%): Students will be required to submit a project proposal outlining the problem they will address, a description of datasets/sources used, and the AI techniques they plan to use.

Dataset Requirements: Students must obtain a reasonable dataset for their chosen topic excluding the datasets used in the lectures and labs. 

The dataset can be obtained from any public data repository and should be referenced. The dataset must meet the following requirements:

• Must have a column containing labels representing the classes identified in the dataset and should contain not less than 4 features.
• Must contain no less than 3 classes.
• Must have at least 100 samples.

COMP40511 Tasks

Task 1

Modify your solution in Task 1 to use a multilayer perceptron (MLP) classifier using scikit-learn with the following parameters: 3 hidden layers, having 30, 15, and 20 neurons respectively. Create a subsection in your report describing the parameters of the MLP model and training method, and how they were set. Also, report the accuracy of the resulting method.

Task 2

. Implement an ensemble model using sci-kit-learn. Create a subsection in your report describing the chosen parameters of the ensemble method, training method, and how they were set. Also, report the accuracy of the resulting method. Compare the results of the ensemble model with the model implemented in task 1 You may also justify the other parameters used in the ensemble model and how it affect the results.

Task 3. 

. Replace the classifier model in your program with a deep convolutional neural network (CNN) with a minimum of 3 convolutional layers using the Keras or TensorFlow library. Create a subsection in your report describing all parameters of the CNN model, training method, and justification of the parameters chosen. Also, report the accuracy of the resulting method and the effects of adjusting different parameters on the results. Include a table in your report comparing the accuracy of the models used in tasks 1 – 3.

Task 4. 

Use a clustering method of your choice to cluster the dataset, and measure how accurately the clusters correspond to the classes in your dataset. Determine the optimum number of clusters. Also, create a subsection in your report where you mention your methods, resulting accuracy, and any other metrics used in the evaluation of clustering results.

III.  Submission Requirements

Your submission must include the following submitted as separate documents:

• Python file(s) (either .py or .ipynb) that contains your source codes (this must not be compressed). Failure to submit this as a separate document will result in your work being returned.
• Report (either in MS Word or PDF) document containing your complete report. This must not be compressed). Failure to submit this as a separate document will result in your work being returned.
• Any supplementary files needed to run the program either as a compressed folder or individually. This includes the dataset used in your work.

The report should have a brief introduction where you summarise what you did. The report should not be more than 3000 words (excluding the cover page, plagiarism declaration, python program, and references). The report should be written in a formal format, including a cover page, a table of contents, an introduction section, implementation, evaluation and references sections. Please note that it is not necessary to write a comprehensive literature review.

When you take a significant portion of code from somewhere, it must be referenced appropriately. You must also reference any scientific paper(s) or other sources that you used for deciding which methods or parameters to use. For this coursework, it should not normally be necessary to take significant portions of code from anywhere except the online documentation of the libraries used.

IV.  Assessment Criteria

Your overall grade can be understood with reference to the table at the end of this document.

V.    Feedback Opportunities 

Formative (Whilst you’re working on thecoursework)
You can ask for guidance during the weekly lab sessions and you will also be given the opportunity to book appointments to discuss the assessment outside of class time.
Summative (After you’ve submitted the coursework)
You will receive specific feedback regarding your coursework submission together with your awarded grade when it is returned to you. Clearly, feedback provided with your coursework is only for developmental purposes so that you can improve for the next assessment or subject- related module.

VI. Resources that may be useful

Referencing styles please use Harvard as detailed here
Guide to planning your time here and an automated planner here
Remember to use resources such as calendars to manage the use of your time between lectures and labs to work on this coursework.

VII. Moderation

The clarity of this specification, and the appropriateness of the assessment criteria, have been checked by two members of the Department of Computer Science. The submissions will be marked by two members of the module team, and the grades awarded will be reviewed by another member of the Department to check for consistency and fairness.

VIII. Aspects for Professional Development

• The ability to demonstrate an in-depth understanding of the theory behind AI and data mining techniques and their critical evaluation.
  
• To critically appraise the main issues involved in real-world AI and data mining systems and the consequences of their use.
  
• To design, implement and critically evaluate machine learning systems to solve real world problems.
  
• The ability to Critically appraise and apply AI technology to extract useful and actionable insight from data and use it to assist decision making.
  
• To demonstrate the ability to write a detailed technical report describing AI solutions.