Ελληνικά
This Έργο is not specific to Greece. For more details, please visit OpenAIRE Explore.
Σύνδεση με ερευνητικά αποτελέσματα.
Παρακαλώ συνδεθείτε πρώτα και δοκιμάστε ξανά.
Ολοκληρωμένο
Self-supervised learning of neural speech representations with dutch archival data
ΧρηματοδότηςNetherlands Organisation for Scientific Research (NWO)
Διάρκεια31 Οκτ 2022 - 05 Οκτ 2025
Κωδικός έργου2022.022
Εντολή ανοιχτής πρόσβασηςΔημοσιεύσεις: ΌχιΕρευνητικά δεδομένα: Όχι
Αναγνωριστικά
Χρονοδιάγραμμα έργου100%
Περιγραφή
Modern artificial intelligence models for automatic speech recognition are produced by big tech, nowadays. In this project we are going to investigate if we can do this on an Academic budget. Specifically, we want to produce a pre-trained model that can be fine-tuned to a speech or speaker recognition model.
Εταίροι (4)
Ξεκινήστε ένα νέο DMP στο Loading
Modern artificial intelligence models for automatic speech recognition are produced by big tech, nowadays. In this project we are going to investigate if we can do this on an Academic budget. Specifically, we want to produce a pre-trained model that can be fine-tuned to a speech or speaker recognition model.
