Case Study DLIC

Deep Learning Industrial Challenge (DLIC 2017)

Here is the link to the official module description:
Modul: Case Study: Modelling, Simulation, Control and Optimization (MSCO) (English)

Kick-off Meeting: 25.10.2017, 11 a.m., room 1518

Course Overview

Allgemeine Informationen, die auch im Vorlesungsverzeichnis (Stundenplan) zu finden sind.
LehreinheitIngenieurwissenschaften
StudiengangAITM Automation and IT Master
Semester3
DozentenBartz-Beielstein, Chandrasekaran
FachCase Study: Deep Learning Industrial Challenge (DLIC 2017)
Raum1518 Steinmüllerallee 6
WochentagMittwoch
Uhrzeit

Nearly all industrial processes and machines are controlled and monitored using sensor data. With the actual trend of Industry 4.0 this reliance on sensor data is even increasing. Recording the sensor data is only half of the story, industry needs specialists to actually process and apply algorithms on this data in order to derive meaningful information. In this project, you will work with sensor data from our industrial partners (see below) to identify events of quality problems in the sensor data.

You have to apply preprocessing methods, select a suitable machine learning (e.g., deep learning, but any other method is feasible) event detection / quality control algorithm and implement your solution.

This case study is organized as a challenge.
During the first part of the study, you will learn about a) typical test cases from our industrial partners and b) standard tools from machine learning. In the second part of DLIC 2017 you have to apply this knowledge. The final goal is to develop an event detector to accurately predict changes in a time series of drinking water composition data. Further information about the current challenge, which is also part of one of the leading conferences in computational intelligence (GECCO 2017), can be found here: http://www.spotseven.de/gecco/gecco-challenge/gecco-challenge-2017/

b. What do you learn?
 State-of- the-art machine learning approaches (deep learning)
 Dealing with practical engineering problems from industry
 Documentation of scientific projects
 Working in teams

c. Requirements
 Scientific working and writing
 Basic programming skills
 Interest in complex real-world problems
 Knowledge of data mining, process control, statistics as introduced in
DDMO/APCO Partners:

d. Project
 Bosch Thermotechnik
 Opitz Consulting
 Thüringer Fernwasserversorgung
 Aggerverband
 IMProvT research project
 OWOS research project
 KOARCH research project

e. Tutors:
T. Bartz-Beielstein, S. Moritz, S. Chandrasekaran, SPOTSeven Lab Team

f. Duration:
Flexible start. Case study can be prepared during the summer months.