National Repository of Grey Literature 2 records found  Search took 0.01 seconds. 
Evolution strategies for policy optimization in transformers
Lorenc, Matyáš ; Neruda, Roman (advisor) ; Pilát, Martin (referee)
We explore the capability of evolution strategies to train a transformer architecture in the reinforcement learning setting. We perform experiments using OpenAI's highly parallelizable evolution strategy and its derivatives utilizing novelty and quality-diversity searches to train Decision Transformer in Humanoid locomotion environment, testing the ability of these black-box optimization techniques to train even such relatively large (com- pared to the previously tested in the literature) and complicated (using a self-attention in addition to fully connected layers) models. The tested algorithms proved to be, in gen- eral, capable of achieving strong results and managed to obtain high-performing agents both from scratch (randomly initialized model) and from a pretrained model. 1
Speciální třídy P-matic v intervalovém prostředí
Lorenc, Matyáš ; Hladík, Milan (advisor) ; Zeman, Peter (referee)
This work focuses on generalizing some easily recognizable subclasses of P-matrices into interval settings, including some results regarding these classes. Those classes are those of B-matrices, doubly B-matrices and BR π -matrices. We derive characterizations, some necessary conditions and sufficient ones, plus we introduce some of their properties, such as are the closure ones and a few conditions the entries of such matrices satisfy. Then we proceed to state a way to generate instances of some of these interval matrix classes. 1

See also: similar author names
1 Lorenc, M.
3 Lorenc, Marek
2 Lorenc, Marián
2 Lorenc, Martin
8 Lorenc, Michal
2 Lorenc, Miroslav
Interested in being notified about new results for this query?
Subscribe to the RSS feed.