Our partner, the National Library, invites you to a “launch seminar” for their newest speech-to-text model for Norwegian.

They have invited the police, Schibsted, and TV2 to share how they use speech-to-text. They have been part of their beta group and will discuss the model’s handling of Norwegian in various contexts.

After lunch, the seminar will become more technically oriented, especially suitable for those interested in deeper insights into the technical and linguistic aspects related to the development of such AI models.

The event is open to all and will be streamed on nb.no. To secure your spot, get your free ticket here!

Program
10:00 Welcome by Aslak Sira Myhre, National Library

Part 1: NB-Whisper and its usage
10:20 Brief introduction to Whisper by Per Egil Kummervold, National Library
10:30 Use of speech-to-text in the police by Bente Skattør, Police
10:50 Applications for speech recognition with NB-Whisper by Anders Haarr, Schibsted
11:10 Use and legal requirements for accessibility by Anne Cathrine Gotaas, TV2
11:30 Lunch break

Part 2: NB-Whisper: technological aspects
12:30 Status of ASR in Norwegian today by Marie Røsok, National Library
12:50 Evaluation and results by Per Egil Kummervold, National Library
13:10 How to build a dataset for training Whisper by Freddy Wetjen and Rolv-Arild Braaten, National Library
13:30 Advanced use of NB-Whisper by Per Egil Kummervold and Javier de la Rosa, National Library
14:00 Questions and discussion