Setup the repo

This commit is contained in:
1jamesthompson1
2025-08-06 10:59:19 +12:00
commit 574ff09929
8 changed files with 28884 additions and 0 deletions

File diff suppressed because it is too large Load Diff

Binary file not shown.

14
readme.md Normal file
View File

@@ -0,0 +1,14 @@
# Seminar notes for AIML430
I am completing Application and implications of AI at Vic uni. However due to being away I am completing it via distance. The seminars are inperson and not recorded. In lieu of attending I will instead complete notes and get them ticked off by the tutor who took the seminar.
Instructions are;
> as per the course outline, you need to "attend" at least ten seminars in order to receive a grade higher than a C in this course. If your remote participation makes this infeasible, we will allow you to instead write summaries of the seminars (in your own words), including a short reflection on the seminar, e.g. what you learned and questions you may have asked if you had attended. These are not graded, but will need to be a "reasonable effort" to count towards this requirement. You may not use any AI tools to help complete this.
Where there are two seminar sets week 5-6 and week 9-10. I only need to attend ten across them all.
## Set 1
- [ ] Mon 4th august (athletic performance and conversational diagnostics)
- [ ] Tue th august (poversity estimates and my semainr)

View File

@@ -0,0 +1,9 @@
# Poverty estimates
This study is investigating using existing survey data and earth observation data to train a ML model that can then predict poverty rates using only earth observation data.
I think in principle the idea of using earth observation and readily available imagery to model systems is very powerful. Mainly as once you havea trained model the only limit on your use of the model throughout the world is based on the models generability.
Some questions that I have are:
- I assume that they used some other "off the shelf" style models or datasets to get the earth observation data. However did they consider/investigate the idea of using satalite imagery directly with Neural Networks?
- With regard to the under performance in the urban areas was it due to the data not being dense enough in that area (i.e trying to predict 10,000 people from the same size set of earth observation as opposed to only 100 people in rural areas), or was it more to do with urban areas have differnet relationships with assests and spending so less applicable of a method?

Binary file not shown.

Binary file not shown.