Summary of Mosel: 950,000 Hours Of Speech Data For Open-source Speech Foundation Model Training on Eu Languages, by Marco Gaido et al.
MOSEL: 950,000 Hours of Speech Data for Open-Source Speech Foundation Model Training on EU Languages
by Marco Gaido, Sara Papi, Luisa Bentivogli, Alessio Brutti, Mauro Cettolo, Roberto Gretter, Marco Matassoni, Mohamed Nabih, Matteo Negri
First submitted to arxiv on: 1 Oct 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Artificial Intelligence (cs.AI); Sound (cs.SD); Audio and Speech Processing (eess.AS)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The paper addresses the lack of fully open-source speech foundation models (SFMs) that comply with regulatory efforts. Existing SFMs fall short due to the absence of publicly available model weights, code, and training data under open-source terms. To fill this gap, the authors collect suitable training data for 24 official EU languages, releasing automatic transcripts for 441k hours of unlabeled data under a permissive CC-BY license. This facilitates the creation of open-source SFMs for EU languages. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper tries to make speech foundation models (SFMs) more open and available. Right now, most SFMs don’t share their model weights, code, or training data in a way that’s fully open-source. To change this, researchers are collecting training data for 24 European Union languages and sharing some of the transcripts they made using this data. This will help create new, open-source SFMs for these languages. |
