Institutional Repository of Vilnius University
Skatinamojo mokymosi modelių tyrimas alternatyviai ir augmentinei komunikacijai /
Abstract
dc:descriptionConventional PECS card selection systems rely on fixed rules and do not adapt to individual children’s needs or to the stochastic behaviour typical of real therapy settings. The aim of this thesis is to identify the most effective combination of a reinforcement learning algorithm and card selection strategy for improving the effectiveness of augmentative and alternative communication using PECS cards. The results show that the PECS card selection strategy is the dominant factor in AAC system effectiveness, outweighing the choice of reinforcement learning algorithm, and that a stochastic child model reveals essential behavioural differences not visible with a deterministic one. Under the conditions of the noisy virtual child profile, similarity heuristics acquire a decisive role: unreliable feedback destabilises learning, while the semantic similarity component helps to compensate for this effect and maintain the effectiveness of the reinforcement learning model.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2026
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Ivanauskaitė, Paulina,
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:ELABAETD308337005&prefLang=en_US