Επαλήθευση συγγραφέα με χρήση προεκπαιδευμένων γλωσσικών μοντέλων
Authorship verification using pre-trained language models
View/
Keywords
Contrastive learning ; BERT ; RoBERTa ; Siamese ; Επαλήθευση συγγραφέα ; Προ-εκπαιδευμένο γλωσσικό μοντέλο ; Ομοιότητα συνημιτόνου ; Embeddings ; Bi-LSTMAbstract
In our everyday life, no one can dispute the necessity of artificial intelligence
applications. These applications cover the largest to the smallest needs of modern
humans. Knowledge, curiosity, security, and recognition are some of the basic human
needs that people seek to satisfy through the internet. Social networking pages, chat, and
blogs provide information and communication. There are also many incidents where we
cannot verify the authenticity of a text as to its author. Authors who have left their mark
in world literature can be easily recognizable. The difficult part is identifying the writer
in the chaos of the worldwide web, and also for documents where the opinions of
linguists and scientists diverge. It is easy to recognize texts from ancient or classical
literature, but it is difficult to recognize in real-time the characteristics of an anonymous
or forged writer. As the internet expands, the production of written language multiplies,
and the field of artificial intelligence and an Author Verification system become
increasingly necessary. It is no coincidence that many private or public enterprises and
university units have integrated this field into their services.
Postgraduate Studies Programme
Τεχνητή Νοημοσύνη - Artificial IntelligenceDepartment
Σχολή Τεχνολογιών Πληροφορικής και Επικοινωνιών. Τμήμα Ψηφιακών ΣυστημάτωνCorporate Department
Εθνικό Κέντρο Έρευνας Φυσικών Επιστημών «Δημόκριτος»Number of pages
88Language
GreekCollections
Except where otherwise noted, this item's license is described as
Αναφορά Δημιουργού-Μη Εμπορική Χρήση-Όχι Παράγωγα Έργα 3.0 Ελλάδα
Αναφορά Δημιουργού-Μη Εμπορική Χρήση-Όχι Παράγωγα Έργα 3.0 Ελλάδα