Projects | AI-FM

This page contains a number of proposals for Bachelor- and Master thesis topics. If you are interested in any of these projects (or in any of our research directions more generally), feel free to send us an email!

Active Measure MDPs

Deciding the value of information in Markov decision processes

Belief Compression

Inexact compression of beliefs for scalable solving of POMDPs

Constrained Multi-agent RL

Reinforcement learning for multiple agents under cost constraints

Delayed Observations

Solving Markov Decision Processes with delayed observations

MDP Discounting

Applying different discounts to Markov Decision Processes

Memory in RL

Memory architectures in reinforcement learning

Offline RL Shielding

Safety in offline reinforcement learning via shields

Offline Causal Discovery

Causal Discover with Online Interaction for Data Augmentation

RNN for uPOMDPs

Verifiable RNN-Based Policies for Uncertain POMDPs

safely transferring reinforcement learning agents under adversarial attacks

ongoing

	Sensitivity-guided Exploration for Learning Markov Chains Mees Meuwissen
	Robust Permisive Policies for Interval Markov Decision Processes Bram Pellen
	Verifying Inter-Process Communication Between ASML Components Renato Feroce
	Improving Deep Q-Learning Performance through Imitation Learning Bas Neeleman
	Card Games as Constrained Reinforcement Learning Problem Michel van Wijk

complete

2023	Active Measuring in Uncertain Environments Merlijn Krale
	Scaling Online Planning for MPOMDPs with Many Agents: New Algorithm Variants and Addressing Scaling Issues Maris Galesloot
	The Underlying Belief Model of Uncertain Partially Observable Markov Decision Processes Eline Bovy
	Safe Reinforcement Learning From Pixel Observations Using a Stochastic Latent Actor-Critic Yannick Hogewind
2022	Optimal Order Execution for FX trading Pleun Koldewijn
2021	Model Learning of Deterministic MDPs Marck van der Vegt
	History-based Rewards for POMDPs Serena Rietbergen
	Optimal Maintenance Strategies for an Industrial Scrubber System Ilse Pool
	Evaluating Adversarial Attack Detectors using Formal Verification Methods Reinier Joosse
	Grouping of Maintenance Actions on Sewer Pipes: Using Deep Reinforcement Learning and Graph Neural Networks David Kerkkamp
	Formal Verification in Uncertain POMDPs Nils Neerhof
2020	Approximating Black-Box Deep Neural Networks using Active Learning as a Proxy Measurement for Robustness Christoph Schmidl
	Entropy-guided decision making in multiple-environment Markov decision processes Marnix Suilen