Machine Learning & Astrophysics Hack Days is a series of 4 seminars organised at PhD level to explore machine learning techniques that are often used in astrophysics during the winter semester 2020 at the ETH Zürich.
The official website for the seminars can be found here: https://astroml-hackdays.org. At this address you will find all the information as videos, links, further readings, etc. In this repository you will find all the code examples and exercises.
The seminars will be completely online, due to the COVID-19 pandemic. They will recorded live from the ETH and streamed live. A recording of the lectures will be available later on this website for everyone to enjoy.
Information on live streaming and videos will be posted on this website as soon as available.
In this folder you will find additional material that will help you in getting enough information on the fundamentals required to understand neural networks and TensorFlow Keras.
Additionally the 2nd edition of the book by Umberto Michelucci and Michela Sperti (Applied Deep Learning) planned for the end of 2021 have an online companion that will grow with time at https://adl.toelt.ai.
The material for the first seminar can be found HERE. There you will find the slides and the Jupyter notebooks for the hands-on sessions.
The material for denoising autoencoders can be found HERE. There you will find the slides and the Jupyter notebooks for the hands-on sessions.
The material for classification can be found HERE. There you will find the slides and the Jupyter notebooks for the hands-on sessions.
The material for classification can be found HERE. There you will find the slides and the Jupyter notebooks for the hands-on sessions.
The seminars have been organised and supported by Prof. Dr. Sascha Quanz, Associate Professor at ETH Zurich, Department of Physics, Exoplanets and Habitability. Logistics and lectures will be organised and given by Dr. Daniel Angerhausen and Umberto Michelucci. Michela Sperti will join the team as teaching assistant.
In case you have questions please don’t hesitate to contact umberto.michelucci@toelt.ai or dangerhau@phys.ethz.ch at any time.