domrand

Reproducing results of Domain Randomization for Transferring Deep Neural Networks from Simulation to the Real World https://arxiv.org/abs/1703.06907

maml_numpy

Implementation of MAML in numpy, deriving gradients and implementing backprop manually

generative_models

Implementations of fundamental deep generative models. Currently: Autoregressive models, VAEs, GANs, and Diffusion models.

emcee_simulation

Gazebo simulation of Utah Robotic Mining Project robot for NASA Robotic Mining Competition from 2016

object_collections

Code to accompany our CoRL 2019 paper

rock_detector

[abandoned] Domain randomization to train a CNN rock detector for NASA Robotic Mining Competition. See https://github.com/matwilso/domrand for more recent work in this direction, with better code

reinforce

Basic implementations of REINFORCE using neural networks. Just to help learn about RL initially. in both raw numpy and tensorflow.

boxLCD

box2D physics with low resolution and binarized rendering. Testbed to enable quick iteration on predictive modeling and robot learning research

carsim2d

Simple top-down 2D car simulator made in C++ using box2D for physics and SFML for rendering.

emcee

matwilso.github.io

webpage

datamining_project

school project for cs5140: data mining at the u of u