-------
 The Zero Resource Speech Challenge 2021 enhanced at NeurIPS 2021'
-------
 

 Dear colleagues,
 
We are pleased to officially announce the second round of the 2021 Zero Resource Speech Challenge, on spoken language modelling, which will take the form of a competition at NeurIPS. While the benchmark remains the same, this enhanced edition of the challenge introduces some twists, including a new, optional, TRACK 2 on visually grounded spoken language modelling.
 
We invite:
 
- New approaches to spoken language models, particularly, but not limited to, larger ('higher-budget') language models
- Systems exploring the utility of weak supervision with images or videos (Track 2)
- Further, substantive development of existing systems doing spoken language modelling, including major modifications to or improvements upon systems submitted in the first, Interspeech round of the challenge
 
For more detailed information, please see the ZeroSpeech 2021 website:
 
https://zerospeech.com/2021/
 
We are also in the planning stages of a NeurIPS workshop on the general theme of unsupervised and semi-supervised speech, which will be an excellent venue to submit detailed reports on your systems. While the workshop submission deadline has yet to be announced, we expect that it will be in mid-October.
 
For groups wishing a little extra time to improve their systems after the workshop deadline, the *final* competition submission deadline to be considered for inclusion in the presentation of the results at NeurIPS has been set for November 19th.
 
Please get in touch with us if you have questions or comments about this continuation of the 2021 Zero Resource Speech Challenge, and we look forward to seeing the results!
 
Best,
The Zero Resource Speech Challenge Organizers
-------