james/AIML430_seminar_notes
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

First draft of all seminar 2 summaries.

Browse Source
This commit is contained in:
1jamesthompson1
2025-09-23 16:21:45 +12:00
parent 240de0e11a
commit 1785424164
3 changed files with 20 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
readme.md
View File

@@ -21,5 +21,5 @@ These ones are in September. Each one is 1 hour long. I completed 6 seminars.
- [x] tue 16th sept stock price prediction team 6 - [x] tue 16th sept stock price prediction team 6
- [x] mon 22nd sept humour team 10 - [x] mon 22nd sept humour team 10
- [x] tue 23rd sept safe guarding infrastructure team 3 - [x] tue 23rd sept safe guarding infrastructure team 3
- [ ] wed 24th sept classroom team 1 (partly done) - [x] wed 24th sept classroom team 1
- [ ] wed 1st oct crop management team 8 - [x] wed 1st oct crop management team 8

10
wed1stoct/wed1stoct Normal file
View File

@@ -0,0 +1,10 @@
Ai in crop management
Team 8 survey a broad range of papers that discuss in depth various aspect and applications of AI in the horticulture space. To summarize I will briefly mention each of the application areas.
The first application is that of pest management. Using AI can help with identifying pests in a cheaper and more precise way than manual methods. Having better monitoring allows for more precision application of the pesticides.
Identifying diseases that are on your crops is vitally important to manage them and either cure them or mitigate damages. Using drone to capture images you can then use the YOLO model to identify and classify troubled plants.
Using XAI methods on ML models can help give farmers real insight into the options and choices that they have. These methods involve using IoT senors to collect large amount of data and modelling this data with respect to crop type. From here you can apply XAI methods on the black box models to understand the effect of the fields environment on crops.
The water use of agriculture and horticulture is shockingly high. Using AI to more precisely and efficiently use the Water could have very high benefits. The idea is to use various easily accessible data sources and applying ML models to predict irrigation needs. These have been shown to be quite effective however still far from off the shelf systems.
Another use of remote sensing is to predict the yield that you will get. This is particularly useful when you rely on natural resources like rainfall instead of human powered irrigation. Relatively simple models like random forests perform with very high accuracy.
Overall there are many different ways that AI can help improve the efficiency and effectiveness of farming. It seems to be an application area that the risks involved with applying AI are the more mundane risks of unintended consequences, sufficiently wrong AI models and slow/inequable adoption.

12
wed24thsept/wed24thsept.md
View File

@@ -1,9 +1,13 @@
AI in the classroom AI in the classroom
AI for students. Help student do work, help student learn with tutoring, simulations as specific form of tutoring. Team 1 topic is AI in the classroom and more broadly how it can be used within the education sector. The two areas of applications can be split into uses for students and uses for teachers.
AI for teachers, plagerism detector, automatic marking, student monitoring, misc topics (admission, assistants etc) Students can use various AI systems which mainly fall into AI tutoring systems and AI powered tools. I see that AI powered tools as an avenue that could impair a students learning by making tasks objectively easier (similar step change to have a basic calculator to a graphing calculator). Whereas the AI tutoring systems seem like an avenue which could level the playing field of education by allowing everyone globally to have a high quality private tutor (conditional on having a sufficiently effective and aligned tutoring system, which is a large conditional). The team point out fair observations in that current tutoring systems are working toward a poor metric of near term test scores. This is a poor objective as test score are a low information feedback and near term is not the goal of education. Furthermore this is exacerbated by profit driven companies being the developer of these products.
Custom leranign content is good. Automated grading is better. Interatice learning and tutors is good (although might not be as effective). There are various ways in which teachers can use AI systems as well. There are quite a wide variety of tasks which AI can help with like grading, lesson preparation etc. Like in the student space many of these tools are being developed with unclear objectives and with for-profit motives. Furthermore as the teachers directed AI systems work on classwide basis (as student directed tool could, but don't have to) they are collecting a large amount of data and have a large influence on the group. This means general privacy and surveillance concerns as well bias propagation are a concern.
Ethical concerns: personal data, change role of teacher It seems like broadly an area that will see massive changes as we move forward. Globally there appears to be underfunding of education and stressed out teachers. Therefore any system offering help (even if bad in the long run) could be adopted. I would hope that more effort is focused on the next generation of minds and so not to let them be overly influenced by the system of the cheapest seller.
Some question that I would of asked if I was there:
- Any more information on the BrainCo headbands, Where was the -presumably- pilot study conducted? Seems like an overly creepy situation.
- The misalignment of the developers incentives seem to of concern. Were there many projects that were from within the education sector (i.e no profit motives)?