Give me a pivot point and I will move the Earth. Who said that?
This week I'm learning modeling in Fusion 360 and trying to find at least 10 minutes to write this dev log. And my pivot point, in this case, is the beginning from where I start telling about my project.
First, I don't like that the term "robot" can mean so many things to people. Some people call "robots" any mechanical moving object. Maybe that object can become a robot if it has a purpose, like an arm that can pick tomatoes. Or maybe just any RC car with a pair of eyes can be a robot? Or anything with AI on board?
I want to convey exactly one thing that I have in mind. I believe it must be all together.
So, this was a quick overview of "Why I didn't finish the project in one year". I keep telling myself that I would finish with the hardware part in October (2022), but I switched to software, to a "real neuron" model. It's in quotes because I put a special meaning there. It might be inspired by neurons, but it must be something different than the current ANN. It might be SNN, but I don't know how many answers researchers have in that field. It's quite a big topic, but I will uncover it eventually. However, I want to focus on small tasks first.
I started with a simple task - designing a face for an LCD display. My robot must have a charismatic face. It must display different emotions. I decided to take some existing model and tweak it a little. And questions started pouring in.
- Do I convert photos of real people into sketches and then into pixels?
- Can I use GAN models? How to make an animation?
- Can I use diffusion models (because they look more artistically pleasant)?
There is still much to learn and discover, but I am excited to see where this project takes me.