DISPLACE-Med (DISPLACE 2026)

Following the grand success of the DISPLACE 2023 and 2024 challenges, which received overwhelming participation and support from both academia and industry, we are thrilled to launch this new version focusing on advancing research in multilingual medical and conversational speech processing.

The event is organized by the Learning and Extraction of Acoustic Patterns (LEAP) Lab, Department of Electrical Engineering, IISc Bangalore, and the Department of Electronics and Communication Engineering, NITK Surathkal, in collaboration with UPES, Dehradun.

We received highly positive feedback and wide recognition for the previous challenges during INTERSPEECH 2024, where the winners were announced on stage. Building on that success, DISPLACE-M 2026 introduces new tracks, tasks, and opportunities for the research community.

Challenge Tracks

Track 1: Speaker Diarization in Multilingual Scenarios

Track 2: Automatic Speech Recognition in Multi-Dialect Settings

Track 3: Topic Identification

Track 4: Text Summarization of Medical Dialogues

Track 1 focuses on speaker diarization (“who spoke when”) in multilingual conversations involving code-mixed speech.Track 2 addresses ASR for multi-dialect speech recorded in noisy medical environments.Track 3 involves topic identification (e.g., maternal health, child nutrition, sanitation).Track 4 focuses on generating concise, clinically relevant summaries from medical audio conversations.

Baseline systems will be released for all tracks, and a leaderboard evaluation platform will be available for participants to track their progress in real time.

Important Links

Registration: https://displace2026.github.io/#registration

Timeline: https://displace2026.github.io/#timeline

Form to Express Interest: https://forms.office.com/r/U5zLg8nihB

Website: https://displace2026.github.io/

Paper can be submitted to the INTERSPEECH 2026 main session.