Call for Demos:

Speech Translation, Speech Synthesis, Text-to-Speech (TTS) and Context-aware Speech Processing at IWSLT 2025

Welcome to the International Conference on Spoken Language Translation (IWSLT) 2025! This year, IWSLT expands its scope to include demonstrations of state-of-the art cross-lingual communications systems and innovative components that support it. We seek to show systems that highlight cutting edge research in Speech Translation, Speech Synthesis, Text-to-Speech (TTS) and Context-aware Speech Processing through an engaging Demo Session.

We invite researchers, practitioners, and industry professionals to showcase innovative systems, tools, and technologies that advance the field of speech translation, speech synthesis, TTS and context-aware speech processing. This session provides a platform for interactive demonstrations, fostering collaboration and knowledge exchange among participants.

Topics of Interest

Submissions are encouraged in, but not limited to, the following areas:

  • Simultaneous Speech-to-Text Translation (English-to-X)
  • Simultaneous Speech-to-Speech Translation (English-to-X)
  • Context-aware Speech-to-Text Translation
  • Multilingual Simultaneous Speech Translation
  • Low resource Speech Translation
  • Adaptive Segmentation Strategies for Simultaneous Speech Translation
  • Model Compression Techniques for Qwen2-Audio
  • Advances in neural TTS systems
  • Multilingual and low-resource TTS solutions
  • Expressive and emotional speech synthesis
  • Audiovisual synthesis techniques
  • TTS applications in spoken language translation systems
  • Real-time and low-latency speech synthesis
  • Voice Conversion
  • Deep and Shallow Contextual Speech Recognition
  • Domain-specific Speech Recognition using LLMs
  • LLM-based fully-formatted Speech Recognition

Submission Guidelines

To participate in the demo session, please submit an extended abstract of up to 2 pages detailing:

  1. The problem addressed and its significance
  2. Technical description of the system or technology
  3. Novelty and contributions
  4. Demo setup and requirements (e.g., hardware, internet, etc.)

Accepted abstracts will be published in the IWSLT 2025 Proceedings.

At least one author of an accepted demo must register for the conference and present their work during the session.

Important Dates

  • Submission Deadline: April 21, 2025
  • Notification of Acceptance: May 15, 2025
  • Demo Session: August 1, 2025

How to Submit

Submit your extended abstract via the softconf submission portal.

The conference submission portal will be posted in March.

For questions or additional information, please contact the Demo Session Chairs Berrak Sisman (JHU, contact: sisman@jhu.edu) and Maximilian Awiszus (Zoom, contact: maximilian.awiszus@zoom.us).

We look forward to your contributions and to exploring the latest innovations in speech translation, speech synthesis, TTS and context-aware speech processing at IWSLT 2025!