Machine Learning: Basic Course and Intense Study GroupOnline Event
https://notes.desy.de/RLpT8_gDSgabhaA8ew7DvQ?edit
Zoom
Powered by Helmholtz AI, IHRS BioSoft, HIDSS4Health and BIF-IGS offer an online course that will introduce you to machine learning. The potential applications of ML and deep learning methods to natural science research are numerous, including detector development, data analysis techniques, and even physical modelling. These techniques will be essential to ensure the highest-quality and impactful scientific results from existing and future experimental works.
This course is conceptualized as a basic course (a three-day event virtually held via Zoom) that will be followed up by an intense study group that can be booked independently of each other. Participation in the intense study group requires the machine-learning knowledge taught in the basic course, though. Detailed information regarding the intense discussion group will also be provided during the course.
Basic Course
The basic course is for you if you do know Python and want to learn the basics of machine learning in a short block course. You may or may not join the intense study group afterwards to dive deeper into advanced topics. The introductory course will be based on scikit-learn and equip the participants with basic tools necessary to begin implementing and using ML methods. It will involve some theoretical descriptions but focus on hands-on exercises and discussions in the tutor group. After this course, learners can embark on classification and regression projects using scikit-learn and should be able to transfer their knowledge to other platforms, such as TensorFlow and PyTorch.
Intense Study Group
The intense study group (starting at the end of April or May) is for you if you want to go into more detail than in the basic course. You will dive into more advanced topics, such as deep learning and reinforcement learning, and will read and discuss recent literature on machine learning techniques and its applications. We currently plan bi-weekly meetings with durations of 1.5 hours each for about half a year, and approximately one day of effort for you on reading pertinent literature and possibly also implementing some models in-between the meetings.
Previous Knowledge
If you want to enrol in this course, we expect you to bring along knowledge of the Python language as taught in our course "Python from Zero to Data Science" (basic Python, Pandas, Matplotlib).
Technical Requirements
The course uses Python 3 and some data analysis packages such as Numpy, Pandas, scikit-learn, and Matplotlib. We rely on you to use a local Python installation. You may use miniconda to get going if you don’t have a working Python installation on your machine. To execute code, you may use jupyter, ipython, plain scripts and a python interactive session, to traverse through the course material.
More details can be found in these instructions.
You can obtain the course material (notebooks and datasets) as well as some installation instructions (in a README.md file) from the learner pack. Download this zip archive to your harddrive and work inside throughout the entire course.
Please send us your corresponding output (via a survey we will send you) so we can check beforehand if everything works. In case you struggle, we do offer a Zoom call for you to double-check your setup prior to the Machine Learning course. However, we strongly recommend that you ask your local IT support first, they know your system and center policies best!
If you struggle with installing anaconda (to use Jupyter and the libraries mentioned above) and do not have access to a Jupyter server at your institution, please consider using Google colab as a fallback. You are free to use other integrated development environments for Python as well, but our support will focus in Jupyter Notebooks and Google colab.
Code of Conduct
We adhere to the Software Carpentries' Code of Conduct. You can report violations to the instructors, helpers, or research school coordinators.
Registration
All interested colleagues are welcome to register. Members of the graduate schools BIF-IGS, IHRS Biosoft and HIDSS4Health will be preferred. To register, you must first log in to the booking system on the top right. If you do not have access to Helmholtz AAI via a Helmholtz centre, you can alternatively use your ORCID ID, which you can create here: https://orcid.org. A quick guide to Helmholtz AAI is available at https://go.fzj.de/HelmholtzAAI.
Central Learner Pad
https://notes.desy.de/RLpT8_gDSgabhaA8ew7DvQ?both