Process
I attended the launch of the Young Coders MeetUp at RedHat where we learnt about machine learning through an awesome website called machine learning for kids, created by Dayle Lane. We all learnt how to create a machine learning models to recognise whether inputted comments were good or bad. We used scratch to animate a robot to show the results. After that we all did our own projects. For me, as I love cats so much it was natural for me to use the image recognition feature to build a machine learning model to recognise cats. I had to upload lots of images of cats from the internet (which I really enjoyed doing) to use as supervised training data. Then I created the scratch program and tested it out.
After this I decided to build a AI pet challenge to submit for the Coolest Projects competition. This was a bit more complex and so I had to contact Dayle through my mum. At my Makerspace I made a square box to put the sample pictures of cats and dogs for spectators to interact with the AI model with on the day of the competition. I particularly enjoyed learning to use the Laser Cutter and engraving some designs onto the side of my box. When I tested it at the Coolest Projects it was quite hard to set up and took a long time to do the training but it worked brilliantly once spectators came in to look. Building this AI pet challenge really helped me to understand more about machine learning, image recognition, datasets and supervised learning. It was really nice to be able to use scratch to build the program, as it's simplicity allowed to lean a lot more about AI.
