C04

C04

Learning fast and efficient hyperparameter control for deep reinforcement learning on small datasets

Project summary

Deep reinforcement learning (RL) algorithms are a powerful class of methods for optimizing sequential decision making and control problems, and are the driver behind many real-world applications. We hypothesize that deep RL applications can considerably benefit from a focus on hyperparameter tuning. Specifically, we aim to develop automated techniques for setting the hyperparameters in a sample-efficient manner for small datasets, thus reducing uncertainty. We will create appropriate benchmarks and then investigate the performance of currently existing hyperparameter optimization (HPO) methods, to then develop automated approaches. This will include a dynamic meta-level control with the use of a robust hyperparameter transfer for online HPO for offline RL and neural architecture search (NAS).

Our methods

  • Data-driven modeling
  • Neural networks
  • Meta-learning
  • Local perspective

Principal investigator 1

Doctoral researcher position 1

(supervised by PI Awad)

Tasks

Firstly, the creation of cost-effective benchmarks using small datasets. Secondly, the investigation and development of automated techniques for determining hyperparameters in deep reinforcement learning algorithms, with a specific emphasis on enhancing performance in small dataset scenarios. Additionally, the development of new, efficient HPO and NAS methods that will enhance accuracy and decrease uncertainty for offline deep RL applications.

Requirements

  • Master’s degree or equivalent in computer science, machine learning, or related fields
  • Solid understanding of and hands-on experience with reinforcement learning and deep learning
  • Experience in hyperparameter optimization techniques is a plus
  • Advanced programming skills in Python (preferably familiarity with PyTorch)

Principal investigator 2

Doctoral researcher position 2

(supervised by PI Bödecker)

Tasks

Analysis of importance and the dynamics of hyperparameters of different RL algorithms, robust, hyperparameter transfer methods, and an evaluation in applications from the biomedical domain.

Requirements

  • Master’s degree or equivalent in computer science, reinforcement learning or related fields
  • Solid understanding of and hands-on experience with reinforcement learning and deep learning
  • Advanced programming skills in Python (preferably familiarity with PyTorch)
  • Strong mathematical background

Administrative Manager

Marc Schumacher

Institute of Medical Biometry and Statistics,
Faculty of Medicine and Medical Center –
University of Freiburg