Evolved Policy Gradients
Interpretable Machine Learning through Teaching
Learning a Hierarchy
Generalizing from Simulation
Meta-Learning for Wrestling
Learning to Model Other Minds
Better Exploration with Parameter Noise
We've found that adding adaptive noise to the parameters of reinforcement learning algorithms frequently boosts performance. This exploration method is simple to implement and very rarely decreases performance, so it's worth trying on any problem.
Learning to Cooperate, Compete, and Communicate
Robots that Learn
We've created a robotics system, trained entirely in simulation and deployed on a physical robot, which can learn a new task after seeing it done once.
Spam Detection in the Physical World
Learning to Communicate
Attacking Machine Learning with Adversarial Examples
This post describes four projects that share a common theme of enhancing or using generative models, a branch of unsupervised learning techniques in machine learning.