Georgian language nlp using word vectorization and RNNs

Author: დავით სოსელია
Co-authors: levan shugliashvili, irakli koberidze, sandro jijavadze, shota amashukeli
Keywords: NLP, Machine Learning, Recurrent Neural Networks, LSTMs, Word2Vec, encoding, Georgian Language
Annotation:

the research explores the use of word vectorization and Recurrent Neural Networks for Georgian language analysis on word, sentence, and paragraph level. Exploring possible modifications such as form random generation as possible remedies to Georgian dataset size. a corpus of a few million words has been gathered for the project and used for evaluation in conjunction with other open-source datasets.