Moses - Machine Learning for Dynamical Systems

Display language

To modulepage Generate PDF

Module / Version:
#50937 / #1

Validity:
Seit WiSe 2020/21

Default display language:
English

Module title:
Machine Learning for Dynamical Systems

Credits:
6

Responsible person:
Ghani, Abdulla

Grading:
benotet

Type of exam:
Mündliche Prüfung

Zugehörigkeit

Faculty:
Fakultät V

Institute:
Institut für Strömungsmechanik und Technische Akustik

Area of expertise:
35312100 FG Datenanalyse und Modellierung turbulenter Strömungen

Prüfungsausschuss:
Physikalische Ingenieurwissenschaft

Kontakt

Office:
MB 1

Contact person:
Ghani, Abdulla

E-mail address:
ghani@tu-berlin.de

Website:

https://www.dmf.tu-berlin.de

Learning Outcomes

Students will be able to code Machine Learning architectures for their purpose and learn to judge on the obtained results. Students will develop understanding on how to use machine learning algorithms to model dynamical systems and a propulsion application.

Content

The program is composed of lectures and hands-on training in groups of 2. Each day is scheduled as follow: lecture during the morning, hands-on session during the afternoon. The first four days are dedicated to introducing various notions on chaotic systems and machine learning. The last day is dedicated to applying these notions to try to model the dynamics of a flame under acoustics forcing. The details of the program are hereunder: Lecture 01 - Introduction to AI and neural networks (NN) Hands-on Session 01 - Introduction to tensorflowTopic Lecture 02 -Introduction to deep learning Hands-on Session 02 - Introduction to keras and feedforward NNTopic, Lecture 03 -Introduction to Recurrent Neural Networks (RNN) Hands-on Session 03 - Intro to RNNTopic Lecture 04 -Introduction to Physics-Informed Neural Network (PINN) Hands-on Session 04 -Introduction to PINNsTopic Lecture 05 -Introduction to Convolutional Neural Network (CNN) Hands-on Session 05 -Intro to CNNTopic: Development of CNNs to identify stars

Module Components

Workload and Credit Points

Course-independent workload:

Workload description	Multiplier	Hours	Total
Course preparation	1.0	30.0h	30.0h
Daily programming exercises	5.0	4.0h	20.0h
Final report	1.0	30.0h	30.0h
Preparation of presentation	1.0	10.0h	10.0h
Project work	5.0	3.0h	15.0h
Time present at the course	5.0	12.0h	60.0h
			165.0h(~6 LP)

The Workload of the module sums up to 165.0 Hours. Therefore the module contains 6 Credits.

Description of Teaching and Learning Methods

This course will be in online form. We will have prepared videos and will discuss on the content during live video sessions using a whiteboard. Students will perform hands-on coding and present their results online.

Requirements for participation and examination

Desirable prerequisites for participation in the courses:

Knowledge in numerical methods and python programming Basic knowledge in fluid dynamics

Mandatory requirements for the module test application:

This module has no requirements.

Module completion

Grading

graded

Type of exam

Oral exam

Language

English

Duration/Extent

about 30min

Duration of the Module

The following number of semesters is estimated for taking and completing the module:
1 Semester.

This module may be commenced in the following semesters:
Winter- und Sommersemester.

Maximum Number of Participants

The maximum capacity of students is 16.

Registration Procedures

Exam registration in accordance with the AllgStuPO

Recommended literature
No recommended literature given

Assigned Degree Programs

Semester:

Zeige auch ausgelaufene Studiengänge und StuPOs

This module is used in the following Degree Programs (new System):

Studiengang / StuPO	StuPOs	Verwendungen	Erste Verwendung	Letzte Verwendung
Computational Engineering Science (Informationstechnik im Maschinenwesen) (M. Sc.)	1	11	SoSe 2021	SoSe 2024
Luft- und Raumfahrttechnik (M. Sc.)	1	3	SoSe 2023	SoSe 2024
Maschinenbau (M. Sc.)	1	4	SoSe 2023	SoSe 2024
Physikalische Ingenieurwissenschaft (M. Sc.)	2	26	SoSe 2021	SoSe 2024
Technomathematik (M. Sc.)	1	9	SoSe 2023	SoSe 2024

Miscellaneous

Please prepare the following frameworks on your personal computer: - Python 3.x version - Tensorflow library Not mandatory, but useful: Jupyter notebooks

Machine Learning for Dynamical Systems