CST4012 Machine Learning for Developers Summative Assessment 2026 | Middlesex University
Looking for Plagiarism free Answers for your US, UK, Singapore college/ university Assignments.
| University | Middlesex University |
| Subject | CST4012 Machine Learning for Developers |
CST4012 Summative Assessment
General Information
The end-to-end use case is a common challenge of information extraction from a population of documents at scale. The goal of the business is to understand two key pieces of information – the payment terms and the limitation of liability. In order to succeed at this goal, we need to build a set of models to not only ingest and read the documents, but to identify the correct context relating to these clauses and parse out a standardized answer. The development of your end-to-end use case will be documented via a design and implementation document. You will record a 7-minute highlight demonstration video to show the main features of the application.
General Project Ideas
There are no limitations to what you can produce for this assessment. However, you should consider the complexity of the application and the timeframe that you have to complete the work. Realistically, you will have 6 weeks to complete the work – the first 5 weeks are likely to be spent understanding basic concepts and techniques and early exploratory work. In that regard, here are some first considerations and tentative steps to get the ball rolling:
- Data must be extracted from the source documentation
- There is a low amount of language variability in the vast majority of the documents
- Understanding the data quality, overall corpus patterns, and common language patterns of these two clauses
- Identifying a relatively small sample of data to annotate that maximises model performance
- Building an effective and high-performing approach to consistently extract
- Turning your results into a script and an easily deployable solution
There is a set of Jupyter notebooks provided from Week 6 onwards, which focus on Points 1.-
4. of the above. These notebooks provide exercises for the following topics:
- Data Gathering and Selection
- Data Ingestion
- Pre-processing and Exploratory data analysis
- Data Understanding and Annotation
- Dataset Curation
- Modelling Approaches
These notebooks build up towards the final end-to-end development use case. However, for this final coursework, you need to go beyond what is provided by these Jupyter notebooks and address Points 5 and 6. above. To ensure that you have a broader understanding of what it takes to build a fully useful solution – not a singular model, you will build an end-to-end script for the ingestion of new files through to the end of creating an output that is usable and valuable for a (less technical) user. This final development is guided by an additional set of exercises provided to you.
The assessment submission is an individual piece of work. However, brainstorming, discussion and exploration of ideas can be carried out in pairs/groups. Please refer to Section 6 of the module handbook with regard to integrity and academic misconduct.
You are welcome to discuss ideas with the tutor and receive feedback on the project before committing to a final solution. More detailed feedback on your work will be provided via the weekly Formative Assessment submissions – see Formative Assessment Specification in the Assessment and Feedback folder on the course page.
Considerations
Time Frame Complexity
- How much time do I have?
- Can the project be completed by the deadline to a high standard?
Complexity
- Do I have or can I learn the required skills in the time that I have?
- Is my implementation plan too ambitious?
Aesthetics
- There are no marks available for aesthetics, so you are encouraged not to spend a lot of time designing and implementing your own assets (unless you have enough time, of course).
Money
- For the end-to-end Use Case, you will need to use Python and Docker
- You will also need to download resources from GitHub.
- All of these resources are free of charge, so you need no money.
Suitability
- What is my target audience – is it usable for them?
- Does my application violate any ethical or moral principles or issues?
Tutorials
You are welcome to use tutorials to assist in the development of your application. Any tutorials that you follow should be documented clearly in your design document. Failure to do so could result in an accusation of plagiarism – please refer to Section 6 of the module handbook with regard to integrity and academic misconduct.
If you decide to start your project using an online tutorial and expand on the tutorial by adding further development, features, etc., these additions must be clearly described in the Design and Implementation Document, and a link to the tutorial must be provided. Applications that follow this approach will achieve lower marks and will not achieve higher than a 2:1 – see page 5 for the marking criteria.
Design and Implementation Document
The Design and Implementation Document does not have a word limit, but should be no longer than 2500 words. If you find that you have exceeded 2500 words, you may be writing too much. The most important thing is that the document covers the main design and implementation points of the project. The content of the Document will be covered in detail during Week 6. Below is an example of the expected sections:
Introduction
- Aims and Objectives
- Concepts
- Challenge of information extraction from a population of documents at scale
Requirements Specification
- End-to-end script for the ingestion of new files
- Usable and valuable for a (less technical) user
ML Development
- Data gathering, selection and ingestion
- EDA
- Data Understanding, Annotation and Curation
- Modelling approaches
Increase the usability of the end-to-end ML pipeline
- Unit tests for data/input/output types and validation, pipeline integrity, error handling, edge cases, etc.
- Logging (for debugging of pipeline)
- Output options (add additional output data options beyond Excel)
- Model Card (for model documentation (features, algorithm, parameters) and evaluation)
- Input options: Provide User Documentation and basic UI for your users (optional)
- Containerise (using Docker) to wrap up and deploy the solution
Evaluation and Conclusion
- What did you do?
- What was successful/unsuccessful?
- Did you encounter any problems?
- What did you learn?
References
- URLs of tutorials
- Any books/articles referenced
- Harvard or IEEE style
NOTE: Work produced for the formative assessment may be incorporated into the summative assessment.
Assessment Weighting
100% of the module.
What to Submit
The submission will be split into two parts: ‘Implementation and Video’ and ‘Report’.
Upload a zipped folder containing the following to the ‘Implementation and Video’ tab:
ML for Developers, End-to-end development Use Case (Implementation)
7-minute Highlight Video Demonstration
In addition, you must upload the Word or PDF file of the Design and Implementation Report.
Where to Submit
Submit your work to the dedicated submission link in the Assessment and Feedback folder on the course page.
Marking
This assessment is worth 100% of the overall mark for CST4012.
Flexible Rates Compatible With Everyone’s Budget
Hire a Professional Essay & Assignment Writer for completing your Academic Assessments
CST4012 Summative Assessment Marking Criteria
| Section/ criteria | 1-4
1st |
5-8
2:1 |
9-12
2:2 |
13-16
3rd |
17-20
Fail |
|---|---|---|---|---|---|
| Design and Implementation Document 30% | Excellent: all stages of design and implementation have been clearly covered. Development in relation to existing code, tutorials, and assets is clearly documented. | Good: most stages of design and implementation have been clearly covered. Most development in relation to existing code, tutorials, and assets is clearly documented. | Satisfactory: some stages of design and implementation have been clearly covered. Some development in relation to existing code, tutorials, and assets is clearly documented. | Minimal: few stages of design and implementation covered, and/or development in relation to existing code, tutorials, and assets are poorly documented. | Very poor: stages of design and implementation are poorly covered, and/or development in relation to existing code, tutorials, and assets is not adequately documented. |
| *Originality 20% Zero marks here for a minimally changed or unchanged tutorial | Evidence of excellent effort at originality. The application has major novel aspects, which are well documented. | Evidence of good effort at originality. The application has some novel aspects that are well documented. | Evidence of adequate effort at originality. The application has a few novel aspects that are well documented. | Evidence of little effort at originality. The application has very few novel aspects, which are well documented. | No evidence of original work, tutorial used with no added functionality, poorly documented. |
| Advanced Features 20% | Excellent evidence of development of/use of advanced features. | Good evidence of the development of/use of advanced features. | Satisfactory evidence of development of/use of advanced features. | Adequate evidence of development of/use of advanced features. | Poor/no evidence of development of/use of advanced features. |
| Effort and Quality of Application 30% | Excellent effort has been put into the development process from design to implementation of a polished application. | Good effort has been put into the development of most stages of the application. | Satisfactory: some effort has been put into development in some stages of the application. | Little effort has been put into the development of the application. | Very little effort has been put into the development of the application. |
A note on Originality
Some words related to originality: novelty, independent, creative, unusual. We are looking for work that was created and developed by you. Try to do something a bit different, even if it has been done before.
Demonstration of Application
Your work must be demonstrated via a video demo submitted with your application and design document. Pay attention to the following requirements:
The video should be no longer than 7 minutes – only the first 7 minutes will be viewed – everything after 7 minutes will be ignored.
The video should highlight the main features of your software, for example, interaction, highlights of the implementation, security features, and attack scenarios.
Details of how to record a video demonstration can be found on the course page in the Assessment and Feedback folder.
Viva Voce
Where cases of plagiarism or overuse of AI tools and/or tutorials are suspected, you may be invited to a viva voce to discuss your work and clarify how the application was developed, and to explain the code and how it works. Use of AI tools and any tutorials referenced will also be discussed, along with how they were applied to the immersive application
Flexible Rates Compatible With Everyone’s Budget
Hire a writer to get plagiarism free assignment answers of this question
Looking for Plagiarism free Answers for your US, UK, Singapore college/ university Assignments.
