cv

Ph.D. in Computer Science and Artificial Intelligence

Sequential Decision Making, TU Delft, NL

Efficient Uncertainty Quantification in Deep Reinforcement Learning. Supervised by Prof. Matthijs T.J. Spaan and Dr. Wendelin Bohmer.

• Part of EU Horizons Project Epistemic AI

Research Assistant

Intelligent Systems, Karlsruhe Research Center of Information Technology, GER

Worked on natural language processing algorithms for requirements management with pretrained models (BERT).

Research Assistant

Cognitive Systems, Karlsruhe Research Center of Information Technology, GER

Developed gradient estimation techniques for trust-region methods in model-based RL.

M.Sc. Mechanical Engineering

Karlsruhe Institute of Technology, Germany

Thesis Title: Model-Based Reinforcement Learning for Constrained Policy Optimization in Robot Locomotion. Supervised by Prof. Marius J. Zoellner

• Graduated with distinction
• Majored in Robotics and Medical Engineering

Visiting Student

Tohoku University, JP

Robotics

B.Sc. Mechanical Engineering

Karlsruhe Institute of Technology, Germany

• Majored in Engineering Design

Abitur

Gymnasium Korntal Muenchingen, Germany

MLSS Scholarship

Machine Learning Summer School, Krakow Poland

Students@Bosch Fellow

Robert Bosch GmbH

GfSE Student Award

German Association for Systems Engineering

DAAD Annual Scholarship

German Academic Exchange Service

Cholesky Ordered Projection Q-learning (COP-Q): Guiding Safety-First Exploitation-Exploration by Multi-Objective Uncertainty

Guopeng Li, Moritz A. Zanger, Matthijs T. J. Spaan and Julian F. P. Kooij

Under review.

Universal Value-Function Uncertainties

Moritz A. Zanger, Max Weltevrede, Yaniv Oren, Pascal R. Van der Vaart, Caroline Horsch, Wendelin Böhmer andMatthijs T. J. Spaan

International Conference on Learning Representations (ICLR).

Contextual Similarity Distillation: Ensemble Uncertainties with a Single Model

Moritz A. Zanger, Pascal R. van der Vaart, Matthijs T. J. Spaan and Wendelin Böhmer

International Conference on Learning Representations (ICLR).

How Ensembles of Distilled Policies Improve Generalisation in Reinforcement Learning

Max Weltevrede, Moritz A. Zanger, Matthijs T. J. Spaan and Wendelin Böhmer

Neural Information Processing Systems (NeurIPS).

Value Improved Actor Critic Algorithms

Yaniv Oren, Moritz A. Zanger, Pascal R. van der Vaart, Matthijs T. J. Spaan and Wendelin Böhmer

Neural Information Processing Systems (NeurIPS).

Diverse Projection Ensembles for Distributional Reinforcement Learning

Moritz A. Zanger, Wendelin Böhmer, and Matthijs T. J. Spaan.

International Conference on Learning Representations (ICLR).

Safe Continuous Control with Constrained Model-Based Policy Optimziation

Moritz A. Zanger, Karam Daaboul, and J. Marius. Zollner.

IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS).