The session provided insights into the current advancements and challenges of speech-to-text technology, and industry leaders and researchers in the field shared their experiences and discussed future directions for this technology. A special focus was on the Norwegian Whisper model, trained by the National Library of Norway. We would like to thank all of our speakers and guests, and hope those who couldn't join us at the event will find the recordings useful.
NB Whisper: "Veien hit og veien videre" - Per Egil Kummervold
Open challenges in automatic speech recognition - Torbjørn Karl Svendsen
Speech technology research - Giampiero Salvi
Norwegian dialect classification by humans and machine - Phoebe Parsons
Semantically Meaningful Metrics for Norwegian ASR Systems - Janine Lizbeth Cabrera Rugayan
Whisper@uio - Dagfinn Bergsager
Transcription and diarization at the NRK - Henrik Lied
Speech to text as backbone for ambient AI - Mikkel Dahl Slettebø, Øystein Nygaard
Tutorial - Finetune ASR models - Sushant Gautam
LLMs for WER Reduction - Sachin Gaur
This event is partly funded by the European Union's DIGITAL EUROPE Programme, under Grant Agreement No. 101083966, and the Norwegian AI Cloud project, funded by Forskningsrådet with Grant Agreement No. 322336.
 |
 |
|---|