AI-driven Testing of Driverless Cars
Driverless cars rely on artificial intelligence (AI) techniques. Unfortunately, AI is prone to biases. In case of driverless cars, this has been witnessed by high profile failures, such as a recent fatal Tesla autopilot crash. Therefore, there is a need to build robustness into driverless car technologies. Robustness entails developing methods that guarantee safety and dependability of learning in various operational scenarios. Testing is one crucial activity to ensure the robustness of an AI-based cars.
To address the challenge of testing AI-based cars, to ensure their robustness and safety, by combining knowledge in software testing and machine learning.
Hands-on machine learning for testing driverless cars, including deep learning and reinforcement learning
Opportunity to publish scientific papers
Bachelors in computer science
Knowledge of machine learning algorithms
Experience with Keras or Tensorflow is an advantage
- Dusica Marijan
- Aizaz Sharif