The learning scenario focuses on machine learning and pattern recognition. It lets students train a machine on recognizing plants.
| Creator | Jan Pawlowski, Martin Idzik |
| Subject | Biology, computer science |
| Length | 45 minutes |
| Pedagogical Approach | Problem based learning |
| Competences | Abstraction Pattern recognition Machine learning |
| Grades | Primary school, grade 5-6 |
| Technologies | Mobile phones / tablets with camera Internet access Computer (if coding the algorithm) https://teachablemachine.withgoogle.com/ |
1. Context setting
Students should discuss how to distinguish plants – what are criteria to distinguish leaves
2. Exploration
Students will do a short walk with the task to find three different trees. They should take pictures on their mobile phones from different perspectives.
Students should formulate the problem (“identifying objects…”)
Students should find criteria how to match a picture with the name of the tree.
3. Elaboration
Two sorts of solutions can be elaborated:
Developing an algorithm in pseudo code. This should define the steps from matching their (self defined) criteria with attributes of a tree (e.g. leave form, color, …)
As a second activity, students can use the teachable machine to train the recognition of different pictures of leaves / trees.
The elaboration includes support by the teacher regarding formulating / coding as well as the use of the teachable machine
4. Production
Students present their solution
In a group review, suggestions for improvements are given
Students then take their trained machine to the same trees to find out whether the AI program works
5. Reflection
Students find further examples where pattern recognition / machine learning is used in everyday life (e.g. driving, face recognition with a mobile phone)
Students also discuss what can happen when training data are manipulated.
