Bc. Ján Petrák

Master's thesis

Black-Box Hyperparameter Tuning for Risk-Constrained Reinforcement Learning Algorithm

Black-Box Hyperparameter Tuning for Risk-Constrained Reinforcement Learning Algorithm
Abstract:
Táto práca poskytuje prehľad novodobých prístupov k optimalizácii hyperparametrov, ako aj prehľad plánovacieho algoritmu s obmedzeným riskom RAlph, vrátane jeho veľkého množstva hyperparameterov. Práca pokračuje uvedením nevyhnutných súčastí potrebných pre optimalizáciu RAlpových hyperparametrov, ako je napríklad definíca cieľovej funkcie. Po krátkom popise implementácie, práca prezentuje výsledky …more
Abstract:
The thesis summarizes modern approaches to hyperparameter tuning, followed by a presentation of a risk-constrained reinforcement learning algorithm RAlph, including its wide range of hyperparameters that substantially influence its performance. The continuation of the thesis comprises the necessary setup for tuning RAlph's hyperparameters, such as the definition of an appropriate objective function …more
 
 
Language used: English
Date on which the thesis was submitted / produced: 14. 12. 2021

Thesis defence

  • Date of defence: 3. 2. 2022
  • Supervisor: RNDr. Petr Novotný, Ph.D.
  • Reader: doc. RNDr. Tomáš Brázdil, Ph.D.

Citation record

Full text of thesis

Contents of on-line thesis archive
Published in Theses:
  • světu
Other ways of accessing the text
Institution archiving the thesis and making it accessible: Masarykova univerzita, Fakulta informatiky

Masaryk University

Faculty of Informatics

Master programme / field:
Computer systems, communication and security / Networks and communication

Theses on a related topic

  • No theses on a related topic available.