Voxcraft-sim
A voxel-based physics simulation for evolutionary robotics implemented in CUDA.
More info:
https://github.com/voxcraft/voxcraft-sim
https://github.com/voxcraft/voxcraft-viz
Xenobots
“It’s a biorobotics platform where we can take advantage of the self-assembling properties of frog skin cells to make useful biological robots.” — Dr. Michael Levin
Xenobots are a new form of robot, designed by computers and realized using frog cells in petri dishes, enabling them to move in a desirable way.
In my previous role in the project, I was responsible for improving the computational simulation for the automated design.
More info:
https://cdorgs.github.io/
https://en.wikipedia.org/wiki/Xenobot
Unreal Engine 5 MassEntity Plugin Demo
A tutorial of how to use the experimental MassEntity System in Unreal Engine 5.0 to create 1000 moving cubes.
More info:
https://www.youtube.com/watch?v=0dwgnH3SoC0
https://github.com/liusida/UnrealMassMovementDemo
Demos for Hinton’s Course
I implemented those demos as practice while taking the course.
The Coursera course ‘Neural Networks for Machine Learning’ is no longer available. However, here is the list of videos on YouTube.
Demos:
https://github.com/liusida/geoffrey-hinton-course-demos
Bonus: a PyCUDA implementation of the Jack’s Car Rental example (4.2) from Sutton’s Reinforcement Learning book.
https://github.com/liusida/JacksCarRental-via-PyCUDA
gym-codecraft
A gym-like environment with a Docker container sandbox for the agent to learn to code.
More info:
https://github.com/liusida/gym-codecraft