machine learning

I am interested in combining design principles with neural network algorithms. Some of my works illustrate machine learning concepts that can run interactively inside a browser.


Please read my blog for more information.


hardmaru

Neural Information Processing Systems (2018)

Recurrent World Models Facilitate Policy Evolution
(Selected for Oral Presentation)

Reinforcement Learning for Improving Agent Design
(Deep RL Workshop)

Deep Learning for Classical Japanese Literature
(ML for Creativity and Design Workshop)

Evolution Strategies
(July 2017 - Nov 2017)

A Visual Guide to Evolution Strategies

Evolving Stable Strategies

ESTool

Google Brain Residency
(June 2016 - June 2017)

Handwriting Generation with Javascript

Teaching Machines to Draw

Hyper Networks

Neural Network Evolution
(July 2015 - May 2016)

Backprop NEAT

Neural Network Evolution Playground

NEAT Regression

Large Image Generation
(March 2016 - April 2016)


Generating Abstract Patterns

Generating Large Images from Latent Vectors

The Frog of CIFAR 10

Artificial Handwriting
(Nov 2015 - Jan 2016)

Fake Kanji
偽漢字

Random Radicals:
A Fake Kanji Experiment

Generative Handwriting

Neural Network Generative Art
(Jun 2015 - Jul 2015)

ōtorography


neurogram


neural network art generation


Creatures avoiding planks
(Apr 2015 - May 2015)

creatures avoiding planks

training from scratch

no planks

Neural Slime Volleyball
(Mar 2015 - Apr 2015)

neural slime volleyball

glass version

evolving from scratch

Inverted Pendulum Balancing
(Dec 2014 - Feb 2015)

PID Controller

Enforced Subpopulations

Conventional Neuroevolution

Miscellaneous

Mixture Density Networks

K-Medoids Clustering

Warhol Image Generator

Data Fitting - SGD

slime agent network

Data Fitting - Conventional Neuroevolution

copyright © ōtoro大トロ.net 2014-2019