1

I've been working on my MSc thesis in creative AI, and chose to research a novel form of game AI, with three seperate strands/use-cases, all of which are novel.

I achieved straight A's in this course, so I'm disappointed I've not been able to achieve everything I hoped for the thesis. In hindsight, this was too far out of scope for a three month thesis period, but here I am.

It's due in just over a month (Oct 2) and I've started writing, because I'm not sure what else to do. There are no extensions available. My supervisor has been on holiday the entire time, as well as my course leaders etc, hence why I'm asking here for advice!

I successfully designed the systems and flows, created the work environment (I had to implement the paper it is based on, which had not yet released its code), and made a rudimentary implementation for one use-case which I'm currently testing. It works but it's not a great implementation - my programming skills are not such that I can fully execute the entire flow, solo, in such a short timeframe.

Is there a way that I can present my design frameworks (which are thoroughly researched and definitely can work, I just ran out of time to write the code) as finished research alongside the part I've finished and tested? I'm comfortable switching from a purely technical focus to a design/HCI one if that would help, and it is permissable as this is a creative computing course.

1
  • 2
    It's not really reasonable to expect someone to write a thesis with absolutely no support. You can ask some questions here, but it won't be sufficient to replace actual supervision. Will your supervisor be available soon so you can get some assistance?
    – Bryan Krause
    Aug 28, 2023 at 16:49

1 Answer 1

2

It's dangerous to give advice when I don't know what the requirements are and I don't know what you're really doing. You should really be having this discussion with your instructor -- hard to believe they just gave you a big project and then disappeared for three months. Still, from what you've told me...

It seems like it's time to rescope. It does not matter if you formally committed to the big project already; scope often changes in research. It's much better to proactively narrow your scope and do a good job on the small problem, than to keep working on the big problem and fail to deliver.

Created the work environment (I had to implement the paper it is based on, which had not yet released its code), and made a rudimentary implementation for one use-case which I'm currently testing

This doesn't sound so bad. You identified an interesting paper, implemented it, changed it substantially so that it would solve a different problem, and tested / evaluated the results. Not bad at all.

Better yet, you still have a month (a third of the time!) left. Personally, I would ditch the rest of it and just dig in here. Either improve the implementation so you get better results (especially if your code will be reviewed), or think about other interesting things you can do with the existing codebase.

Is there a way that I can present my design frameworks (which are thoroughly researched and definitely can work, I just ran out of time to write the code)

Maybe? Saying "I have this idea, but no real proof it will work" is not really convincing. I would certainly avoid calling it "finished research." On the other hand, if you have already put a lot of effort into this design, there probably is a way you can include this work. For example, you could motivate the actual work by saying it is a proof-of-concept for your more creative idea. Or in the "future work" section, you could describe the big project as a natural next step in the small project's evolution.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .