27

May

Teaser image to Improving Motion Prediction in Autonomous Driving with Expert Knowledge - A Bayesian Deep Learning Approach

Improving Motion Prediction in Autonomous Driving with Expert Knowledge - A Bayesian Deep Learning Approach

Nadja Klein, TU Dortmund

   27.05.2024

   4:30 pm - 6:00 pm

   LMU Munich, Ludwigstr. 28 VG/II; Room 211b

Autonomous driving is one of the most highly anticipated yet elusive mobility innovations. The field has made significant advances through deep learning, especially in perception and motion prediction. Still, the field faces open challenges, since safety requirements in autonomous driving demand for robust domain adaptations between locations and well-calibrated uncertainty estimates for the numerous high-risk edge cases in urban environments.

To meet those demands, we explore the potential of Bayesian deep learning methods to motion prediction. We demonstrate how a Bayesian approach can be used to regularize commonly employed motion prediction models by utilizing prior expert knowledge. More specifically, we adapt a CoverNet baseline model with a compute-efficient last layer Gaussian Process approximation and integrate prior drivability knowledge. Doing so improves both, robustness and calibration.

Organized by:

LMU Center for Advanced Management Studies

LMU Munich School of Management


Related

Link to From Bach to Natural Machines: Algorithms as Shapers of Music

Lecture Series  •  06.06.2024  •  Große Aula, LMU Munich, Geschwister-Scholl-Platz 1, 80539 München

From Bach to Natural Machines: Algorithms as Shapers of Music

Dan Tepfer, world-class pianist, composer, and astrophysicist, will perform on Thursday, June 6th, in the main auditorium (Große Aula) of the LMU main building.


Link to Some Recent Advances in Exceptional Model Mining: Tales of Potatos, Boris Johnson, and Atrial Fibrillation

Lecture  •  16.04.2024  •  LMU Main Building, Room M001 Geschwister-Scholl-Platz 1, 80539 München

Some Recent Advances in Exceptional Model Mining: Tales of Potatos, Boris Johnson, and Atrial Fibrillation

This MCML workshop covers the topic of Exceptional Model Mining (EMM). EMM finds exceptional subgroups in data by partitioning columns for candidate subgroup definition …


Link to Increasing transparency through preregistration and/or multiverse analyses

Lecture  •  14.03.2024  •  LMU Munich and Zoom

Increasing transparency through preregistration and/or multiverse analyses

LMU Open Science Center invites to a lecture on multiverse analyses by Prof. Dr. Andrea Hildebrandt on March, 14.


Link to HNU-Management-Forum with our PI Albrecht Schmidt

Workshop  •  12.03.2024  •  Hochschule Neu Ulm, Germany

HNU-Management-Forum with our PI Albrecht Schmidt

Lecture of our PI Albrecht Schmidt at the HNU-Management Forum. He will discuss AI, its meaningful use, and the challenges associated with it.


Link to MCML PI Matthias Niessner about Fotorealistic 3D Avatars

Lecture  •  06.12.2023  •  Deutsches Museum Munich and Livestream

MCML PI Matthias Niessner about Fotorealistic 3D Avatars

TUM Professor and MCML PI Matthias Niessner will discuss the creation of photorealistic 3D replicas of the real world.