COMP9491 - Applied Artificial Intelligence

3 minute read medium

Difficulty: 4/10 · Time Commitment: 5/10 · Enjoyability: 9/10 · Mark: 90

Summary

I did COMP9491 (Applied Artificial Intelligence) in Term 2 2023. This course is research focused and consists of components similar to an Honours year, however completed within a group of 3. There are strict prerequisites for this course due to the research focus and it involves finding a research topic within AI and building a project around it within 10 weeks. Our team’s final report can be found on the posts section of this website titled “Automated fake news detection through contextual similarity comparison”.

Positives

  • Very unique course with a heavy research focus. One of the only opportunities to conduct research within a course before doing a thesis.
  • Tutors and lecturers are knowledgeable in their field and are assigned based on your project, helping provide guidance and ideas throughout the term.
  • Abstract course which allows students to research any topic within artificial intelligence and build knowledge within a new domain.

Negatives

  • This isn’t something that the course can change but the course is really fast paced and there are only 10 weeks in the term, making it hard to build something meaningful.
  • Your group can have a huge impact on your course experience. Make sure to find a group within the first lecture based on your interests and any potential pathways you want to explore.

In-depth

Structure

COMP9491 consists of a project proposal, literature review, final presentation and final report. These assessments are similar to those in the honours thesis. The goal of the course is to

  1. Come up with a research topic within the broad field of artificial intelligence
  2. Conduct research to establish the current state of the art within the field and how that your topic is novel
  3. Develop your topic over the majority of the term
  4. Present your findings and results in a presentation and report

Proposal

The proposal is due very early in the term and a lot of emphasis is put on how well you have defined your project. This is really hard to do early on so I recommend starting within the first week and drafting a proposal as early as possible. Brainstorming ideas early on will make your experience much better as having a useful project idea will go a long way.

Literature Review

This is an opportunity to present all changes and improvements made from feedback after the proposal. For my team, our idea was good but it lacked direction and a final goal so we spent the weeks between the proposal and literature review reading through many academic papers and reviewing the state of the art until we were able to come up with a more concrete set of goals and method to achieve them.

Since the literature review is only a short presentation, the main focus is to discuss the most important current research that is the foundation for your work. Learning how to find relevant research papers and gather useful information from them is an essential skill that you develop throughout this course and really useful if you take an honours year or masters afterwards.

Final Assessments

After finishing the literature review, the remainder of the course is dedicated to finishing your project and evaluating its performance. This leads to your final assessments which are a presentation summarising your findings (and a possible demo if time permits) and a final report which goes over every major feature of your work including the background, dataset, algorithms, models, evaluation, and limitations. It is essential to start early as these can be very time consuming and require a lot of preparation and feedback.

Advice

  1. Make the most of your tutor’s advice. One of the best parts of this course is that tutors are assigned based on your project. Consequently, your tutor will most likely be able to give you a lot of guidance and help you understand different concepts that you might be struggling with. Make the most of their advice and write down all feedback they give you.
  2. Find a group with relevant skillsets. AI is a broad term and different groups want to work on complex projects within image recognition, natural language processing, generative AI etc. Make sure that you find a team that wants to work on similar topics as you so that you can all contribute effectively and are all onboard with the idea. This involves taking some time before your first lab and coming up with (1) your own interests and anything you would not want to work on and (2) brainstorming your own project ideas that could be used as a starting point for your group to come up with your final idea.

Updated:

Comments