Loading
0000003522 20W 6SWS PR Advanced Practical Course - Application Challenges for Machine Learning on IBM Power Architecture (IN2128, IN2106, IN212810)   Hilfe Logo

LV - Detailansicht

Wichtigste Meldungen anzeigenMeldungsfenster schließen
Allgemeine Angaben
Advanced Practical Course - Application Challenges for Machine Learning on IBM Power Architecture (IN2128, IN2106, IN212810) 
0000003522
practical training
6
Winter semester 2020/21
Informatics 17 - Chair of Information Systems and Business Process Management (Prof. Rinderle-Ma)
(Contact information)
Details
Allocations: 1 
Angaben zur Abhaltung
Artificial intelligence, machine learning and artificial neural networks are the growth areas of our time. Self-learning algorithms, trained systems and semi-autonomous data evaluation enable companies to benefit from their company data like never before.

The steady increase in data growth combined with a shorter time to market, especially for software products, requires more and more effective and efficient data analysis. In order to make this possible, not only advanced AI frameworks are required, but also modern hardware architectures that perform such analyses. For example, support with high memory bandwidths and GPU accelerators.

With its Visual Insights product, IBM offers a comprehensive range of coordinated software and hardware that offer and support solutions for specifically such emerging scenarios.

If you are interested in the results of last year's course, you may also have a look at our homepage:
https://openpower.ucc.in.tum.de/research/teaching-and-practical-courses/
- Basics of programming
- Mathematical / statistical basics
- Python and R programming skills are a plus
- Knowledge of machine learning frameworks (TensorFlow, PyTorch, etc.) is nice to have
- Bringing your own notebook is a plus
The aim of the practical course is to develop further knowledge in the field of machine learning and its practical applications.

We want to take a look at ML frameworks with you and enable you to solve specific application problems from the business domain on our internal IBM hardware.

In this context, practical tasks are solved using the IBM Visual Insights Framework as well as open source frameworks such as TensorFlow and PyTorch. The applications are implemented on the basis of the Power9 architecture, which, in comparison to the x86 architecture, enables not only the higher memory bandwidth but also a high-performance connection of NVIDIA V100 GPUs via NVLink.

The concepts covered in the course include:
Data Preparation, Hardware Acceleration, Distributed Deep Learning, Large Model Support, Elastic Distributed Training, Auto Machine Learning (H2O Driverless AI), Auto Deep Learning (IBM Visual Insights), Data Science in Enterprise Environments, etc.
  • German
  • English
workshop
Initially, block courses 2 days in 2 weeks
After that, group work

The Moodle platform is used for communication during the course
Details
Für die Anmeldung zur Teilnahme müssen Sie sich in TUMonline als Studierende*r identifizieren.
Note: To register for participation, you must identify yourself as a student in TUMonline. Bachelor students are generally admitted to the practikum, but if there is a large number of students registering, Master students are preferred when allocating places.

Note: Registration takes place via the new matching system. You can find information on this at: http://docmatching.in.tum.de/.

It is also necessary to fill out a form. You can find it at: https://forms.gle/dSjX4BTE473PKGc1A

Access is activated until July 15th, 2020.
Zusatzinformationen
IBM PowerAI: Deep Learning Unleashed on IBM Power Systems Server
http://www.redbooks.ibm.com/redbooks/pdfs/sg248409.pdf

AI and Big Data: on IBM Power Systems Servers
http://www.redbooks.ibm.com/redbooks/pdfs/sg248435.pdf

IBM Power System AC922: Introduction and Technical Overview
https://www.redbooks.ibm.com/redpapers/pdfs/redp5472.pdf
Online information
course documents
e-learning course (moodle)
Additional information
additional information
If you have any questions, please contact Mr. Simon Fuchs (s.t.fuchs@tum.de) or Mr. Omar Shouman (o.shouman@tum.de).

Since corona-related distance rules will probably continue to apply in the coming winter semester (2020/2021), the course will be held online via Zoom in all probability. Only in the event of a complete withdrawal of all distance rules (e.g. due to the availability of a vaccine) we will hold the course in attendance. The link to the course will be announced in good time.

Please fill out the following form. You can find it at
https://forms.gle/dSjX4BTE473PKGc1A
Access is activated until July 15, 2020 inclusive.