But I’m still in the planning stages, mainly figuring out costs, whether to do it as a team project, or learn the hardware myself. I’m more of a software person, making my own Machine Learning systems that’s; (1. Simpler to implement than deep learning, (2. Rights some of the wrongs inherent in decision trees.
But I like the idea of human-like ( not just humanoid ) robots, but increasingly finding my software operating like a child unable to speak. I use that analogy, as they can do simple reasoning: You looked up corn, you also looked up meal. Aha! You must be looking for corn meal. ( Unable to speak isn’t entirely true, I’m just having trouble with rejoinder type input at the moment. )
I also use SMEG: Standardized Minimalist English Grammar, and in the middle of finding an equivalent for French and Japanese later. Which is designed to make natural language more lifelike, and less character-like.
Certain things taking time: I’m probably going to need to wait for Rest API to become refined, but I’m pleased I could theoretically code a robot in Ruby.
I’m considering a simple embodied intelligence, just to learn how to do it.
What I’ve completed: a system of growing an AI system, rather than direct programming. Rather than programming it, you answer a few questions, and it works out how to program a sub routine on its own. It can also do simple restore from backup. Its now in an early version. I don’t know if it will be compatible with Poppy though.
But there is no non hardware limit to how many subroutines it can have. In this way, its more of a seed software.
Edit: Forgot to show off the specific projects:
https://github.com/LWFlouisa/Saasagi-Automatic # This aims to create a system that can carry much of its development automatically.
https://github.com/LWFlouisa/Saasagi-App # This is the main Saasagi app.
https://github.com/LWFlouisa/Saasagi-Subroutines # Building new subroutines all the time, and eventually Saasagi itself will eventually take over.
Additional projects ( These are more subroutines ):
https://github.com/LWFlouisa/PSIFER # Natural language framework using a dynamic and perpetual decision tree.
https://github.com/LWFlouisa/Online-Offline-Dictionary-With-DDG A dynamic and growing dictionary you can use for online and offline. But mainly offline.
Just finished an auto cloner subroutine.