Safe and Secure Deep Learning for Autonomous Driving

Bachelor thesis, Research assistant, Master thesis, Thesis

Research focus: Machine Learning, Mobility, Safe and Intelligent Vehicles

Job Description

Autonomous driving has made huge advances in the last couple of years. However it remains questionable how safe and secure it is. Since autonomous vehicles rely heavily on deep learning, we need to understand how neural networks make decisions and also guarantee that they function robustly.

At the FZI, we are working on robust and explainable deep learning. Our research aims at finding methods to protect autonomous vehicles from both unintended (i.e. malfunctioning) and intended (i.e. hacking) accidents. We are seeking candidates with a strong interest in performing cutting edge research in a very active and exciting area. Possible topics include interpretability, verification and testing of neural networks as well as adversarial attacks and defenses with a focus on autonomous driving.

 

Your Responsibilities

  • Literature research and evaluation of the state of the art
  • Implementation of the proposed approaches in Python
  •  Development of new techniques to increase robustness of deep learning components

Our Offer

  • An interdisciplinary working environment with partners from science, industry and society
  • Insights into cutting-edge research
  • Opportunity to try out the developed algorithms during test-driving on public roads
  • Hardware to train the algorithms
  • A pleasant working atmosphere
  • Constructive teamwork

 

Your Profile

  • Good Python programming skills
  • Knowledge of deep learning
  • Experience with deep learning frameworks (e.g. Tensorflow)
  • Self-reliant thinking and working
  • Motivation and commitment
  • Fluent in English or German

 

Application

We are looking forward to receiving your PDF application. Please send your application containing the following documents:

  • Current transcript of records
  • Curriculum vitae in tabular form

 

Job Description

Start: as soon as possible