UCLA Astronomy Machine Learning Reading Group - Fall 2018

The goals of the reading group are:

• Discuss machine learning terminology and common machine learning methods
• Read about the use of machine learning in different context, both in astronomy and other fields
• Learn to use the common tools that have been developed for machine learning and deep learning (e.g. Scikit-learn, TensorFlow). By doing it this way, we hope that everyone can maximally transfer their knowledge to their own science or work outside academia.
• Learn to use a centralized scientific computing system. We will be using SciServer at Johns Hopkins University. They have generously allowed us to use their system for free for our work. This means that all you'll need is a computer with a web browser to access and run all the machine learning code we'll be doing.

No prior knowledge of machine learning is necessary to join. It is helpful to have some experience with python and the command line.

Readings

Our main readings will be mainly from:

• Hands-On Machine Learning with Scikit-Learn and TensorFlow
  • Safari Books Online: http://proquest.safaribooksonline.com/book/programming/9781491962282
  • github page: https://github.com/ageron/handson-ml
• Deep Learning by Goodfellow, I.; Bengio, Y.; Courville, A.
  • website: http://www.deeplearningbook.org/

Other useful sources:

• Python Data Science Handbook by VanderPlas, J.
  • website: https://github.com/jakevdp/PythonDataScienceHandbook
• Data Analysis: A Bayesian Tutorial - Sivia, D. & Skilling, J. - Not part of regular reading, but this is a useful reference for Bayesian statistics
• Fundamentals of Machine Learning for Predictive Data Analytics - Kelleher, J.; Namee, B.; D'Arcy, A.

Quick-Getting started with SciServer Compute

SciServer.org is a computational cloud environment that Johns Hopkins University (IDIES group) has generously allowed us to use for our projects. Jupyter notebooks/terminal are the interfaces to access datasets such as SDSS. Here's how to clone this github repo into your SciServer account:

1. Create an account at sciserver.org and go to Compute
2. Create a new container (Docker container), choose the type to be Python, and a container Jupyter notebook interface will be created.
3. On the right hand side, go to New-> terminal and a black terminal interface will appear.
4. In this order in the terminal, type each of these at a time (separated by a comma) and hit enter:
5. ls, cd home, ls, cd idies, cd workspace, cd persistent, ls, git clone url-of-git-repository-here
6. the respository will now be in your folder.

Schedule

Meetings will take place on Fridays at 1-3 PM in PAB-3-703. The first hour will be discussions while in the second hour, we'll work on projects. Room changes will be sent via email.

Organizers: Tuan Do (@followthesheep), Bernie Randles (@brandles)

Date	Topic	Readings	Presenter
2018-10-12	Introductions	Intro to Machine Learning	Tuan & Bernie
2018-10-19
2018-10-26