Introduction to Machine Learning Workshop

ADACS Computational Training Gives Future Astronomers the Tools to Map the Stars

Designed to equip astronomers with an overview of the various machine learning (ML) algorithms and introduce tools to create their own machine learning programs, Introduction to Machine Learning workshop was hugely popular and registration was reopened to allow more participants to sign up.

With observatories and satellites recording ever-greater amounts
of data, ML has the potential of becoming a vital tool in processing the huge data input from astronomy facilities, e.g. by automating the process of data classification. The Introduction to Machine Learning workshop provided face-to-face and hands on instructions on building machine learning algorithms. The material covered supervised and unsupervised machine learning models, principal component analysis and artificial neural networks and also featured domain specific data to work with.

Being able to develop efficient machine learning algorithms allows experts to spend more time solving the problems in our skies, and saves resources such as computational power and money by significantly cutting down the processing time for data. It also allows a tighter turn-around for projects, allowing more work to be done in a shorter timeframe.

Astronomers and ML experts worked in collaboration to compile the workshop material using the open-source scikit-learn library in Python and data from the
Sloan Digital Sky Survey and the Galaxy Zoo project. This provided participants with coding examples to copy and build upon for their own work, and also a readily- accessible resource to work with after the annual meeting.

Participants praised the ability to work hands-on with code, alongside the workshop notes. Some participants said they had already begun using the materials and presentation topics to work with machine learning possibilities in their fields. The workshop and resources were very well- received, with participants asking for a more specialised, levelled workshop in future, to expand upon the topics covered at the meeting.

The workshop was held at the 2017 annual meeting of the Astronomical Society of Australia.

An intensive four-hour workshop providing astronomers with industry- applicable machine learning skills.

Provided open-source, professional-level code examples for Python.

The workshop was prepared in collaboration between machine learning experts in the Curtin Institute for Computation and ADACS astronomers

The material is made available on the ADACS github: Australia/adacs-ml-workshop


agreed taught skills useful for their work

would recommend the workshop

Related posts

Dedicated Astronomy Access to the NeCTAR Cloud

by ADACS Learning Resources
3 years ago

ADACS internship report -Zixian Wang (Purmortal)

by ADACS Learning Resources
2 months ago

Introduction to computing and data science for astronomers

by Lewis Lakerink
7 years ago
Exit mobile version