Ensai Documentation

This code uses convolutional neural networks (with tensorflow) to estimate the parameters of strong gravitational lenses. Unfortunately we’re not very good at coding, so you’ll find that the code is messy, not well documented, and crazily written. However, that shouldn’t discourage you from trying it out. Because it’s a pretty pretty cool thing: The code can recover the parameters of gravitational lenses in a fraction of a second. Something that used to take hundreds of hours!

The results of this work have been published in a Nature letter “Fast Automated Analysis of Strong Gravitational Lenses with Convolutional Neural Networks” (Hezaveh, Perreault Levasseur, Marshall, 2017) and another paper, “Uncertainties in Parameters Estimated with Neural Networks: Application to Strong Gravitational Lensing”, submitted to the Astrophysical Journal Letters (Perreault Levasseur, Hezaveh, and Wechsler, 2017). In the next few months we’ll be slowly making this code more user-friendly and extend it to more interesting and complex lensing configurations.

• Installation
• Overview of the code

Citation:

If you use this code for your research please cite these two papers:

1. “Fast Automated Analysis of Strong Gravitational Lenses with Convolutional Neural Networks” (Hezaveh, Perreault Levasseur, Marshall, Nature, 2017)
2. “Uncertainties in Parameters Estimated with Neural Networks: Application to Strong Gravitational Lensing” (Perreault Levasseur, Hezaveh, and Wechsler, ApJL submitted, 2017)