TensorFlow is an open-source software library for Machine Intelligence. In this talk, we will learn how to use it to build and train a neural network with the goal of correctly identifying asteroids in astrophotography data. The dataset used will be from the Sloan Digital Sky Survey, one of the most ambitious and influential surveys in the history of astronomy.
Using this data, we will learn how to create and featurize a training dataset, build and fit a neural network, and train our model to correctly identify asteroids visible from Earth.
This talk is for a wide range of Python developers, from those who have heard of machine learning, but have never experimented with it, to those who have significant experience with neural networks, but have never used TensorFlow before.
The audience should have some basic Python knowledge, but no formal or informal experience with machine intelligence is assumed.
After watching this talk, the audience should know how to determine and develop features, how and why to build a training dataset, how to build and train a neural network, and some other approaches to supervised learning.