2018-11-27 Machine Learning with R at LRZ (HMLR1W18)
| Date: | Tuesday, November 27, 2018, 9:00 - 18:00 |
|---|---|
| Location: | LRZ Building, Garching/Munich, Boltzmannstr. 1, Kursraum 2, H.U.010 |
| Contents: | The statistical programming language R offers a large amount of add-on packages that implement ideas and methods from the field of machine learning. The mlr package (Machine Learning in R) provides a unified interface to many of these algorithms. It allows to utilize supervised methods like classification, regression and survival analyses along with their corresponding optimization and evaluation methods, as well unsupervised methods like clustering. This includes a consistent infrastructure for common data pre- and post-processing steps, feature selection, hyperparameter optimization, resampling and model comparison. As these tasks become computationally expensive, parallelization is desirable and, for high performance cluster systems, conveniently facilitated by the batchtools package. In this full-day workshop, we will provide an introduction to the mlr package and demonstrate how it can be used to deploy highly parallelized machine learning tasks on LRZ high performance computing systems, using the batchtools package. Participants will be able to work on instructor-led excercises and hands-on examples. Preliminary Schedule 09:00 - 10:00: Overview of LRZ Supercomputing & Machine Learning Infrastructure 13:00 - 15:00: mlr - Demo and Excercises |
| Prerequisites: | This course is addressing participants who have good knowledge of R and who are familiar with basic machine learning concepts. While a short overview of the LRZ infrastructure will be provided, prior experience of working with GNU/Linux on remote systems (SSH) is a requirement (consider signing up for Introduction to the LRZ Supercomputing & Machine Learning Infrastructure on October, 8th which covers these contents). All participants are expected to bring their own laptops. |
| Language: | English |
| Instructors: | Janek Thomas (LMU), Johannes Albert-von der Gönna (LRZ) |
| Registration: | Via the LRZ registration form. Please choose HMLR1W18. |
| Material: | If you attended this course, please don't forget to rate it: https://survey.lrz.de/index.php/256376 |