Βαθιά Μάθηση και εφαρμογή της σε εργασίες γλωσσικής τεχνολογίας
Date Issued
December 3, 2019
Type
Πτυχιακή Εργασία
Abstract
Deep Learning is a very young field of artificial intelligence based on artificial neural networks. Artificial neural networks have unique capabilities that enable Deep learning models to solve tasks that machine learning models could never solve. In this dissertation, are presented in great detail the structure elements of deep learning together with some of the most important architectures. In Addition, a deep learning model is created by using the Keras API and Tensor flow frame work, which is made to recognize hand written digits from pictures. The experimental part expands to the field of linguistic sciences, because of the reference to the text translation models and evaluation of them by the NLP metrics WER, METEOR, TER. Last but not least, there is a presentation of word embeddings together with code that replaces the traditional way of calculating them by using pre trained embeddings in Greek.
Subjects
File(s)![Thumbnail Image]()
Loading...
Name
ΠΕ_ΠΛΗ_ΚΙΒΡΑΚΙΔΗΣ_ΙΩΑΝΝΗΣ.pdf
Description
Πτυχιακή Εργασία
Size
2.62 MB
Format
Adobe PDF
Checksum
(MD5):87e6745e38d91cbfd99deb6aac7632ed