Coursework for COMP4132 24-25

Overview

The coursework aims to make use of the machine learning techniques learned from this module to solve a practical problem. The coursework consists of three key parts: 1) main submission: coursework report and code, 2) Individual report, 3) presentation.

Please read this document carefully to see the requirement of the coursework and the explanations and further details.

Important dates

•    Team formation: 26th Nov 2024

•    Final submission deadline: 23th Dec 2024

•    Presentations: 24th Dec 2024

Copying Code and Plagiarism

You may freely copy and adapt any of the code samples provided in the lab exercises or lectures. You may freely copy code samples from the PyTorch documentation, which have many examples explaining how to do specific tasks. This coursework assumes that you will do so and doing so is apart of the coursework. You are therefore not passing someone else’s code off as your own, thus doing so does not count as plagiarism.

You can, and you should look at other code/papers online, but you need to reference any source/material that you have used as inspiration, and highlight what’s your contribution. Turnitin/JPlag will detect any use of external sources automatically. Successful completion means that you are able to explain your solution during the presentations. The university takes plagiarism extremely seriously and this can result in getting 0 for the group coursework, the entire module, or potentially much worse.

Getting Help

You MAY ask the module convenor for help in understanding the group coursework requirements if they  are not clear (i.e. what you need to achieve). Talk to me during the labs, after the lecture, or post your questions on Moodle. Any necessary clarifications will then be added to the Moodle page or posted on the discussion forum so that everyone can see them. You may NOT get help from anybody else (other  than your group mates) to actually do the coursework (i.e. how to do it), including the module convenor.

Task Specification

The aim of this coursework is to offer you an opportunity to put your hands on designing/developing an advanced machine learning based solution. Language Models (LM) are very popular nowadays in the area of Natural Language Processing. In this coursework you are asked to use LM together with other machine learning techniques learned in this module to solve the problem of joke generation. In particular, you are expected to provide a machine learning based solution that is able to generate a joke given an input such as a few starting words, similar to one of your lab exercises but much more compherensive. Your solution should beat least satisfying the following requirement:

•    Be able to generate a full joke which may consists of several sentences.

•    The  generated joke should at least make some sense compared with say random generation, based on the data you used.

•    Additional functionalities that make the solution better.

Note that you may not have enough resources (i.e. GPUs) to perform a thorough training. You can useless data for the training in your laptop or you can use online resources such as Google Colab. You can use the lab as a starting point but your solution should not be the same as the labsolution. You should usethis datasetas the training data.

Team Formation Instructions

You should form groups of at most three students. The group contributions on the coursework will be assessed, but also each individual effort. Each group should select one person as the Team leader. The team  leader  will  be  in  charge  of  organising   meetings,  team  coordination,  group  submissions  and communications.  During the  lab  session  on  26th   Nov,  you  will  form.  your group. You can talk to the convenor for better understanding of the coursework.

Coursework Report (main submission)

The report must be clearly presented in English with no more than 4000 words, excluding all the figures and tables, summarizing how the task is done, justification on your decisions involved, and the results of your analysis. This report should be submitted (with code) via Moodle by the due date. The folder should be named by the group id (to be assigned).

The submission should cover the following:

1.   Team ID (assigned by the module convenor), student names and Ids.

2.    Introduction presents the aims, the problem you solved, and outlines the solution.

3.    Methodology focuses on the reasoning for the certain techniques and designs that you used in this coursework. This describes and explains your chosen methods and design. It is very important to elaborate why you design your solution in your way.

4.    Result and discussion contains a description and analysis of the results.

5.    Conclusion allows you to have the final say on the issues you have  raised in this coursework, synthesize your thoughts, demonstrate the rationale of your solution.

6.    Reference if there are links, papers, codes are used.

Code

All the code, and documentation files should be submitted with the main submission via Moodle. You do not need to upload the model if it is too big. Instead you can put in on online storage such as one drive and baidu disk and put the link together with the documentation.

The code should contain a README file explaining all the included materials, and references to other codes/papers you have used for inspiration. Moreover, some brief documentation(s) should be provided for each code file to explain the code structure and describe how to use the code and data.

Individual Report

Each member of a team is expected to submit a two-page report including the following:

1.   The student information including full name, email, and ID.

2.    A  table  of  participation  marks:  this   should  provide  marks  to   show   how  group  members contributed/collaborated to the coursework. This table should have three columns, 1) student full name, 2) a mark out of 10 and 3) one (or maximum two) sentence(s) for making justification.

3.   A brief explanation of the individual role in the coursework outlining the offered contributions (Maximum one page).

4.   A discussion of individual understandings, findings, and reflections on the coursework and team- working (Maximum one page).

This report should be submitted by each student via Moodle, a separate submission. The format should be:

“Student Number. Pdf” (e.g., 20029784.pdf)

This will be used to assess the role of each individual in the coursework. This report also might be used to ask relevant questions during the presentation.

Presentation

All the teams will present orally 24th Dec. For this, all group members will be required to be present to

deliver a presentation and answer questions from attendees and module convenor. The presentations are open for all the students.

All oral sessions will follow a similar structure to how they are held in atypical physical conference format.

•    The module convenor will introduce each group.

•    The authors will deliver the  presentation (8 minutes maximum) for the audience. Please make sure you practice this before the session.

•    Once the  presentation  has concluded, the module convenor will facilitate a live Q&A period (3 minutes approx.) with the audience and the module convenor.

•    Process repeats for each subsequent group in the session.

Presentation tips

•    Please use a template to make your presentation slides, it will be available on Moodle.

•    Suggest having maximum 8 slides, 1 slide a minute.

•    You need to upload your presentation file on Moodle, (one day) before the presentation.

•    Start off with a brief introduction of yourselves and the key focus of your coursework.

o This should outline the contributions of each student in the coursework.

•    The outline of the presentation should be similar to the structure of the coursework  (e.g. introduction and motivation, methodology, experimental set-up, results, and conclusions).

•    It is a 8-min presentation, do not aim to show every single aspect of your coursework, focus on the most important things/findings. Key points: Make sure you state clearly your motivation and how good your solution is.

•    Be ready for any question and discuss your contribution. You can prepare additional slides/files for any potential questions you expect. You maybe asked to explain your solution, and even show your code, so please make sure that one appointed member of the group can share the screen and show the coursework and the code.

•    All members should contribute during the presentation.

Coursework Marking Criteria

•    Individual Report (10 marks): Each member of the team is expected to submit an individual report.

o Active role

     Does the student actively participate in the coursework?

     What are the student’s contributions to the coursework? How they are relevant and valuable?

o Understanding

     Does  the  report  show  a  fair  student’s   reflection  and   understanding  of  the coursework?

       Does the report clearly highlight the key findings of the student in the coursework?

•    Main Submission (70 marks): Each group should submit report of their machine learning solution  together with the code produced.

o Introduction

     Do the team understand the overall aims and the problem to be solved in this coursework?

     Do they provide a clear description of the solution provided?

o Design/Methodology

     Explanation of the methodology.

     Justification of the proposed methodology..

o Experiments and results

     Are the experiments well designed to test the proposed solutions?

     Do the results support the original idea?

     Is the analysis coherent?

o Writing

     Clear description, reproducibility

     Quality of visual elements, illustrations, tables.

     Quality of References

o Code and data - software quality

     Efficiency, clarity of the code to solve the problem

     Documentation

•    Group Presentation (20 marks): Each group is asked to deliver a 8-min presentation summarising  their contribution + 3 mins for Q&A.

o Quality and clarity of the presentation

o Response to questions from the module convenor and public

o Understanding of their solution

o Individual  participation  in  the  presentation.  All  members  are  expected  to  participate equally.