Munich Datageeks - June Edition

Important: Please bring a valid ID (Personalausweiß) or passport to show security when entering the building.

It is time for our next Meetup. In June, Airbus will host for the first time.

There are two important changes from our usual procedure:

1. Airbus will provide shuttle buses to and from the area.
   The buses will leave 19:00 from Ostbahnhof (Friedenstraße 17a, near Burger King and Kultfabrik; backside of the Ostbahnhof)
   We'll put signs up the buses so that they can be identified easily.
   Two buses will drive back to the Ostbahnhof: First one at 21.30, second at 22.30.
   You can, of course, travel by car (follow the Haupteingang/Besuchereingang signs) or Sbahn (Taufkirchen, W.-Messerschmitt-Str.).

2. For regulatory reasons, you must state your nationality when entering the Airbus area. But of course everybody (!) who registered is welcome, this is only a formality.

Format:

• 2 presentations (each ca. 30-40 min incl. discussion)

• Of course time for networking + food + drinks before, in between and especially after the presentations

• Talks are held in English

The line up:

Christine Kroll - Smart Video Analytics for Unmanned Aerial System Applications

Abstract:

Unmanned aerial systems are equipped with imaging sensors whose data needs to be analyzed automatically in order to reduce the workload of the human operators and to enable autonomous mission functions. Thereby one typical task is the automatic localization, classification and tracking of dedicated objects of interest for which machine learning techniques like deep learning are employed. This talk gives an insight into the challenges we faced when applying such technologies to airborne systems, showing several smart video analytic applications that are developed by Airbus and addresses the question of how to verify and validate these algorithms

Bio:

Christine joined Airbus in 2008 by conducting a dual course of study in Information Technology and since 2011 she is working as a system engineer for Intelligence, Security and Reconnaissance (ISR) missions. In parallel she did a Master of Science in computer science at the Technical University of Munich (TUM) with a focus on Computer Vision and Machine Learning. In her work in the ISR Mission Systems team she is mainly working on smart video analytics with a focus on deep learning. (In her free time Christine likes traveling, swimming, literature and artificial intelligence.)

Janek Thomas - Applying Model-Based Optimization to Hyperparameter Optimization in Machine Learning

Abstract:

The talk will cover the main components of sequential model-based optimization algorithms.
Algorithms of this kind represent the state-of-the-art for expensive black-box optimization problems and are getting increasingly popular for hyper-parameter optimization of machine learning algorithms, especially on larger datasets.
The talk will cover the main components of sequential model-based optimization algorithms, e.g., surrogate regression models like Gaussian processes or random forests, initialization phase and point acquisition.
In a second part, some recent extensions with regard to parallel point acquisition and multi-criteria optimization will be covered.
The talk will finish with a brief overview of open questions and challenges.

--------------------------------------------------------------------------------

Please note that we need to provide a list of meetup usernames to Airbus to check at the entrance. If you do not agree that this information is given to Airbus, please do not RSVP.