My Solution for the Galaxy Zoo Challenge — Sander Dieleman

me@ryanorban.com (Ryan Orban) — Tue, 08 Apr 2014 00:00:00 +0000

Summary

The Galaxy Zoo Kaggle challenge asked competitors to classify galaxy morphologies from telescope images — the same task that hundreds of thousands of human volunteers had done for the original Galaxy Zoo citizen science project. Sander Dieleman won by using convolutional neural networks, achieving a mean squared error that exceeded human annotation consistency on the test set.

Convolutional-Neural-Networks on Ryan Orban

My Solution for the Galaxy Zoo Challenge — Sander Dieleman

Summary