Collective Intelligence for Deep Learning: A Survey of Recent Developments| 大トロ ・ Machine Learning
EvoJAX: A Hardware-Accelerated Neuroevolution| 大トロ ・ Machine Learning
Permutation-Invariant Neural Networks for Reinforcement Learning| 大トロ ・ Machine Learning
Modern Evolution Strategies for Creativity:Fitting Concrete Images and Abstract Concepts| 大トロ ・ Machine Learning
Neuroevolution of Self-Interpretable Agents| 大トロ ・ Machine Learning
Rather than hardcoding forward prediction, we try to get agents to learn that they need to predict the future. Redirecting to learningtopredict.github.io, where the article resides.| 大トロ
We search for neural network architectures that can already perform various tasks even when they use random weight values. Redirecting to weightagnostic.github.io, where the article resides.| 大トロ
PlaNet learns a world model from image inputs only and successfully leverages it for planning in latent space. Redirecting to planetrl.github.io, where the article resides.| 大トロ
Little dude rewarded for having little legs. Redirecting to designrl.github.io, where the article resides.| 大トロ
GitHub In this article I will give step-by-step instructions for reproducing the experiments in the World Models article (pdf). The reference TensorFlow implementation is on GitHub. Other people have implemented World Models independently. There is an implementation in Keras that reproduces part of the CarRacing-v0 experiment. There is also another project in PyTorch that attempts to apply this model on OpenAI Retro Sonic environments. For general discussion about the World Models article, th...| 大トロ
Can agents learn inside of their own dreams? Redirecting to worldmodels.github.io, where the article resides.| 大トロ
Going for a ride. GitHub In the previous article, I have described a few evolution strategies (ES) algorithms that can optimise the parameters of a function without the need to explicitly calculate gradients. These algorithms can be applied to reinforcement learning (RL) problems to help find a suitable set of model parameters for a neural network agent. In this article, I will explore applying ES to some of these RL problems, and also highlight methods we can use to find policies that are mo...| 大トロ
A Visual Guide to Evolution Strategies| 大トロ ・ Machine Learning
Latent space interpolation of various vector drawings produced by sketch-rnn. GitHub This is an updated version of my article, cross-posted on the Google Research Blog. Instructions on using the sketch-rnn model is available at Google Brain Magenta Project. Link to our paper, “A Neural Representation of Sketch Drawings”. This article has also been translated to Simplified Chinese. Introduction Vector drawings produced by our model. Recently, there have been major advancements in generativ...| 大トロ
Recurrent Neural Network Tutorial for Artists| 大トロ ・ Machine Learning