Introduction

(More details in Research)

Objectives

SCRIBE’s primary objective is to develop a Norwegian speech-to-text transcription system for multi-party conversations in realistic recording conditions.

In order to attain the project goal, research and technology development beyond the state-of-the-art is needed within several key areas. These include language universal issues, as well as issues specifically related to the Norwegian language.

Secondary Objectives

  • We will develop models that are robust to disfluencies that are typical in spontaneous conversational speech, that can cope with turn taking and take advantage of the context in the dialog.
  • The models will also support the use of spoken dialects and different orthographies (Bokmål, Nynorsk, or dialect specific).
  • We will define evaluation metrics that predict the quality of the transcription based on semantics rather than merely word error rate.
  • Finally, we will contribute to the theoretical and methodological development of machine learning with sparse data.

We are grateful for funding from The Research Council of Norway within the IKTPLUSS initiative.

News

5. Nov. 2024

SCRIBE members are presenting at ASTIN 2024 in Trondheim, Norway!

1. Sep. 2024

SCRIBE members are presenting at Interspeech 2024 in Kos, Greece!

18. Jun. 2024

Simen Dymbe is joining the SCRIBE team!! He will start in August.

3. Jun. 2024

SCRIBE members are presenting at Fonetik 2024 in Stockholm, Sweden!

22. May. 2024

SCRIBE members are presenting at LREC/Coling in Torino, Italy!

14. Mar. 2024

NordTrans press conference at NTNU on broadcast news transcription in Norwegian. Register here!!

12. Mar. 2024

The SCRIBE team submitted 5 papers to Interspeech 2024!

15. Feb. 2024

Nationalbiblioteket launches the newest Norwegian Whisper model.

18. Dec. 2023

Giampiero Salvi attends the Workshop on Training and Evaluation Data for Italian/Multilingual LLM.

... see all News

Industrial Partners