My Ph.D. dissertation on "Unsupervised sound separation" is online ideals.illinois.edu/items/127515 😎 If you are interested in self-supervised, multi-modal, efficient and federated learning aspects for sound separation, feel free to take a look 👀

The CHiME Challenge is accepting proposals for tasks. Submitting one is easy (organizing the task less so😅) and any idea concerning speech processing is welcomed (e.g. even audiovisual). Proposals may be merged so it is a great opportunity for collaboration !

#SANE2023 is in just over 1 month, Thu 10/26 at NYU in Brooklyn! Talk details are now up for Kyunghyun Cho, Yuan Gong, Anna Huang, Wenwu Wang, and (New!) Gaël Richard. There's still time to register, but we're nearing capacity. Poster deadline: 9/30 Please RT🤗 saneworkshop.org/sane2023/

Fantastic new development — a new journal focused on data-related questions in ML research, backed by JMLR brand!

What if #pyannote could do separation too? 🔀 I am hiring a postdoc to help me with that! 👩‍🎓 Apply here: emploi.cnrs.fr/Offres/CDD/UMR… … or please RT 🙏 #diarization #pyannote

Applications for 2024 MERL Speech & Audio team internships are now open. Come work with us on sound event/anomaly detection, multimodal representation learning, multimodal scene interaction, and audio source separation, towards publication in top venues. merl.com/internship/ope…

This is a great work from my colleague Zhong-Qiu Wang. IMO it is the first work that does unsupervised **fully neural** (beamforming not required) speech separation in a multi-channel setting without resorting to mixtures of mixtures. Performance is close to supervised PIT.

Hi all, We want to announce the new Interspeech satellite workshop called SynData4GenAI (Synthetic Data's Transformative Role in Foundational Speech Models). Please check the details! syndata4genai.org INTERSPEECH 2024

👀

Is it only my impression or is the quality (last three years) of Interspeech reviews (on average) been much worse than other conferences ? Maybe I am unlucky but every year I get crazy stuff which seldom happens elsewhere. Is the reviewers pool so much different from ICASSP ?

This is a very cool work on speech separation in real world settings (AMI meeting). It is robust because it is weakly supervised (diarization labels only) and thus can be trained on real world data, unlike most neural methods which have to rely on synthetic data.

Hi all, This is the third call for papers about the SynData4GenAI workshop. Good news! While the submission data was originally due on June 18th, we'll extend it to June 24th. Please submit your papers at syndata4genai.org We look forward to your submissions!

If you are interested in generalizable speech enhancement that can tackle "speech-in-the-wild" data, different sampling rates and is able to restore audio from different distortions check this out. We have a new challenge at NeurIPS 2024. Website: urgent-challenge.github.io/urgent2024/tim…

I'm excited to announce WAVLab | @CarnegieMellon's XEUS - an SSL speech encoder that covers over 4000+ languages! XEUS is trained on over 1 million hours of speech. It outperforms both MMS 1B and w2v-BERT v2 2.0 on many tasks. We're releasing the code, checkpoints, and our 4000+ lang. data! 🧵

We release a detailed paper, model weights (model and codec) and streaming inference for Moshi! Beyond the model itself, we believe our findings will be useful to audio language models. "Inner Monologue" for the win!