Abstract

Despite rapid progress in deep-learning–driven perception, there still isn’t a complete, full-stack neuro-symbolic ADAS in the field, one that coherently connects sensing, mapping, scenario understanding, decision support, and validation under a shared “world model.” As a result, many deployed systems remain difficult to interpret, brittle under distribution shifts, and expensive to assure against long-tail driving situations.

In this talk, I present a “knowledge-at-the-wheel” vision for moving from isolated AI components to an integrated neuro-symbolic stack for driving assistance. The core idea is to use knowledge representations as the glue between neural perception and downstream decisions: turning raw sensor streams into structured scene descriptions; anchoring them in a semantic HD map as a persistent, queryable world model, and connecting both to scenarios, operational constraints, and safety requirements. Building on this foundation, I outline how symbolic constraints and ontologies can provide explicit invariants and explanations, while probabilistic reasoning handles uncertainty and ambiguity in real traffic. Finally, I discuss emerging approaches for automatically synthesizing probabilistic reasoning programs from domain knowledge and data, reducing manual rule engineering and enabling more traceable, testable, and updatable decision logic. The goal is a practical pathway toward ADAS that is more robust, more explainable, and more verifiable - without abandoning the strengths of modern neural components.

About Current Trends in Computer Science

This lecture is part of the Current Trends in Computer Science Lecture Series by the TU Wien Informatics Doctoral School, where renowned Guest Professors hold public lectures every semester. If you are studying with us, the lecture series can be credited as an elective course for students of master programs of computer science: 195.072 Current Trends in Computer Science. Additionally, you can join courses held by this year’s Guest Professors of our doctoral colleges and the TU Wien Informatics Doctoral School.

Curious about our other news? Subscribe to our news feed, calendar, or newsletter, or follow us on social media.