FoMLAS 2018 - International Workshop on Formal methods for ML-based autonomous systems (FoMLAS)

After the well-known DARPA Urban challenge, there have been significant improvements towards autonomous driving. In the past few years, the major theme when building self-driving cars has shifted to deep learning and probabilistic techniques. When these new algorithms act as key components in autonomous driving, they create substantial technological challenges in terms of explainability (e.g., can I explain what is happening inside the machine-learning algorithm?), predictability (e.g., can I predict what will happen next in the algorithm, or how good can the machine learning component generalize?), and accountability (e.g., when an accident occurs, can one find the root cause, or who is the one to blame?). This goal of this workshop is to facilitate discussion regarding how formal methods can be used to increase predictability, explainability, and accountability of autonomous systems. The workshop welcomes results from concept formulation (by connecting these concepts with existing research topics in logic and games), as well as algorithms, tools or concrete case studies.
Organisers: Chih-Hong Cheng , Indranil Saha