Institutional Repository of Vilnius University
Rekomendacinių sistemų grįstų transformerių neuroniniais tinklais tyrimas /
Abstract
dc:descriptionIn this paper, we focus on recommender systems, and more specifically on those implemented by transformer neural networks. The paper gives an overview of the types of recommender systems, their advantages and disadvantages. The architecture of transformer neural networks and its components such as: positional encoding, encoder, decoder, self-attention mechanism, multi-head attention, self-attention, direct propagation neural networks and input and input mappings are introduced. The evaluation indicators of recommender systems are described, how they are calculated and what are the advantages and disadvantages. Two consistent recommender systems are reviewed: SASRec and BERT4Rec. The datasets used by the reviewed guideline systems are described. A study was carried out with the reviewed recommender systems to test different types of hyperparameters from their performance in the models. The performance of the models was also compared with four datasets.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2023
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Kochanovskis, Jaroslav,
Rights
dc:rights- Statement dc:rights
-
- info:eu-repo/semantics/openAccess
- Language dc:language
- lit
Identifiers
dc:identifier.*- Repository record dc:identifier
- https://repository.vu.lt/VU:ELABAETD192846689&prefLang=en_US