AI Audio WG meeting 05.20.26

Updated May 27, 2026 Audio AI

May 20, 2026

Invited Venkat Rayudu mallik@syntiant.com abigoppong@gmail.com jmarconi@embedur.com Pong Trairatvorakul EAIF Team Calendar artur.balogh@silabs.com Pete Bernard raashid.ansari@silabs.com tomer.badug@ceva-ip.com dan@atym.io Rosina Haberl eiman.kanjo@ntu.ac.uk elia.shenberger@ceva-ip.com erics@embedur.com e.kanjo@imperial.ac.uk venkata.rayudu@silabs.com seunghun.shin@murata.com jon@femtosense.ai

Attachments EDGE AI Audio WG meeting Notes by Gemini

Meeting records Transcript 

Summary

Meeting organizers coordinated event attendance and finalized collaboration strategies for current white paper initiatives.

London Event Coordination
Participants confirmed attendance status for the upcoming London event. Discount codes were offered to facilitate additional registrations.

White Paper Development
Drafting progress continues with new content additions and planned diagram integration. Members agreed to delegate specific sections to ensure comprehensive document coverage.

Project Integration
The open source DPDFNET model was highlighted for potential integration. Publication is delayed beyond the London event to allow for thorough review.

Rate this Summary: Helpful or Not Helpful

Decisions

We’ve updated the Decisions section using your feedback.

Let us know what you think: Helpful or Not Helpful

ALIGNED

  • White paper collaborative review workflow The white paper production process is established to include a collaborative group review phase and encourages all members to contribute a section to ensure content depth and completeness.

Next steps

  • unchecked[Venkat Rayudu] Submit Figure: Add the pending figure to the draft by the end of the day.
  • unchecked[Abigail Oppong] Review Text: Evaluate the speech to text paragraphs added by Venkat and provide necessary feedback.
  • unchecked[Abigail Oppong] Send Update: Send an email summarizing the white paper status including completed sections and pending improvements.
  • unchecked[Abigail Oppong] Discuss Diagrams: Coordinate with Jonathan regarding the inclusion of diagrams in the white paper.
  • unchecked[Artur Balogh] Review Biometrics: Assess the speech biometrics materials prepared by Tomer and provide contributions.
  • unchecked[Venkat Rayudu] Coordinate Contribution: Discuss with R22 regarding potential contributions to the white paper.
  • unchecked[The group] Request Discounts: Contact Rosina Haberl to obtain discount codes for the London event if potential attendees are identified.

Details

  • London Event Logistics and Attendance: Rosina Haberl discusses preparations for the London event, confirming that Frank and Elia will attend. Tomer Badug states they will be on a personal vacation in Thailand during the event (00:00:00). Regarding further attendance, Artur Balogh and Tamashia confirm they are registered for the event, while Venkat Rayudu states they will not be joining (00:08:41). Rosina Haberl offers to provide discount codes for individuals looking to join the event and invites the group to contact them if they know others nearby who might be interested in attending (00:09:28).
  • White Paper Progress and Content Updates: Venkat Rayudu provides an update on the white paper, noting they met with Tomer Badug and Abigail Oppong to add tables and a speech-to-text section (00:01:27). Venkat Rayudu has added two paragraphs to the main draft discussing the similarities between Automatic Speech Recognition and speech-to-text (00:03:59). They are currently working on a figure, which is expected to be completed by the end of the day (00:01:27).
  • White Paper Review and Workflow: Abigail Oppong confirms that their section of the paper is complete and suggests focusing on refining the sectional observations (00:02:27). Tomer Badug suggests that the group should wrap up the document and then have members perform a review to identify areas needing elaboration or clarification (00:03:59). Abigail Oppong notes that Jonathan, who was unable to attend the call, expressed interest in including diagrams in the paper (00:05:22).
  • White Paper Contributions: Tomer Badug expresses a desire for all group members to contribute a small part or task to the white paper. Artur Balogh agrees to contribute and is assigned to review the speech biometrics section, which Tomer Badug is currently working on but has not yet completed (00:06:34). Tomer Badug agrees to transfer their current work on speech biometrics to Artur Balogh for further integration and elaboration (00:07:41).
  • Audio Technology Initiatives: Tomer Badug shares that they recently published a white paper regarding a speech enhancement model named DPDFNET. This model is open source, features a GitHub repository, and is being integrated into the Sonic Scale system to evaluate its performance. Tomer Badug encourages the team to contact them if they have their own projects they wish to spotlight within the group (00:09:28).
  • Future Scheduling and Publication Timing: Regarding the publication of the white paper, Tomer Badug notes it is unlikely to be ready by the London event. Rosina Haberl mentions that the foundation is planning an event in Taipei for late October or early November, although exact dates are not yet confirmed. Tomer Badug intends to send an email with a status update on the paper’s progress and suggests potentially presenting the work at a future event (00:05:22) (00:10:51). As Tomer Badug will be absent for the next meeting, Rosina Haberl takes responsibility for managing the scheduling of that session (00:11:53).