[AUDITORY] [CfP] Zero-Shot TTS and Personalized Speech Enhancement Challenge at GenDA 2025 (ICASSP Satellite Workshop)

["Kim, Minje" <minje@xxxxxxxxxxxx>]

(Apologies for cross-posting)

Dear Colleagues,

We are excited to invite you to participate in the "Zero-Shot TTS at Personalized Speech Enhancement" challenge, which will be held as part of the ICASSP 2025 Satellite Workshop, “Generative Data Augmentation for Real-World Signal Processing Applications” (GenDA 2025). We are hosting a dedicated challenge that bridges the gap between generative data augmentation and personalized speech enhancement (PSE). The challenge focuses on using zero-shot text-to-speech (TTS) techniques to generate high-quality, speaker-specific data that can boost the performance of PSE models. Participants will be asked to:
	1.	Augment Personalized Data with Zero-Shot TTS: Build zero-shot TTS systems using just a short enrollment utterance (~3 seconds) to produce speech samples that capture the target speaker’s unique characteristics.
	2.	Train Personalized Speech Enhancement Models: Use these synthesized samples to train or finetune PSE models that can better handle noisy, real-world speech conditions while preserving speaker identity. Participants can start from the checkpoints provided by the organizers. 

Why Participate?
	•	Benchmark Your Methods: Test your models against baseline systems, established metrics, and other top research teams.
	•	Push the Frontiers of Personalization: Investigate how generative speech augmentation influences downstream speech enhancement tasks.
	•	Shape Future Directions: Contribute insights to an emerging field at the nexus of generative AI and real-world signal processing.

Important Dates:
	•	Challenge Submission: Deadline March 12, 2025
	•	Results will be announced in early April, before the event.
	•	Workshop Date: TBD (one full day before or after ICASSP 2025 main tracks)

Submission & Participation Details:
	•	Detailed challenge rules, datasets, and baseline code are available on the challenge page of the workshop website: https://sites.google.com/view/genda2025/pse
	•	Submission link for both workshop papers and challenge entries: ICASSP 2025 CMT Submission Portal (https://cmt3.research.microsoft.com/ICASSP2025/)
	•	Choose “Satellite Workshop: Generative Data Augmentation for Real-World Signal Processing Applications”
	•	Select “Challenge: Zero-Shot TTS and Personalized Speech Enhancement” as the primary subject area.

The GenDA 2025 Workshop:
The workshop will feature keynotes, technical sessions, and a panel discussion. Accepted challenge participants will have the opportunity to present their work, discuss their approaches, and receive feedback from leading experts and peers. In addition to the Zero-Shot TTS for PSE challenge, another challenge on room acoustics will be announced soon. 

Contact:
For inquiries, please contact:
	•	Minje Kim (minje@xxxxxxxxxxxx)
	•	Jaesung Bae (jb82@xxxxxxxxxxxx)

Join us at GenDA 2025 to explore the transformative potential of generative data augmentation and shape the future of generative AI and signal processing research. We look forward to your contributions and participation!

Warm regards,

GenDA 2025 Organizing Committee
 - Minje Kim, University of Illinois at Urbana-Champaign / Amazon Lab126
 - Dinesh Manocha, University of Maryland
 - John Hershey, Google Research
 - Trausti Kristjansson, Amazon Lab126
 - Jaesung Bae (UIUC; Task Captain of the Zero-Shot TTS and PSE Challenge)
 - Jackie Lin (UIUC; Task Captain of the Room Acoustics and Speaker Distance Estimation Challenge)