@inproceedings {pub3543pub3592,
	title = {Learning Fluid Flows},
	author = {Theodoros Georgiou AND Sebastian Schmitt AND Markus Olhofer AND Thomas B{\"a}ck AND Yu Liu AND Michael Lew},
	year = {2018},
	month = {July},
	abstract = {Computational FLuid Dynamics simulations are able to produce very complex and large outputs that describe the physical properties of fluids in various domains, such as flow of air around an object, the flow of air and benzine in an internal combustion engine and so on. These simulations are often too complex to be analyzed properly in a manual process. Data reduction and feature extraction techniques are usually used in order to simplify the simulation output and make it
more comprehensible. With the increasing number of simulations as well as their complexity, there is a need of automated processes that can analyze these complex outputs. Inspired by the success of CNNs in Computer Vision we design a novel CNN architecture tailored to the data produced by CFD simulations. We show the capabilities of CNNs in capturing and processing flow patterns. In this paper we try many different architectures and compare them on different tasks that depend on flow patterns. },
	publisher = {IEEE},
	booktitle = {IEEE World Congress on Computational Intelligence}
}