Poly-SIM Grand Challenge at ACM MM 2026

Deadline: May 10

Participate in Poly-SIM 2026, a speaker ID challenge at ACM MM 2026, testing AI models on audio and face data with language switch and face removal.

Join the Poly-SIM 2026 challenge, pushing speaker ID beyond perfect data. Train on face + English audio, then test your model on:

• Audio only (face removed)
• Language switch (English to Urdu)

Can your system still recognize the speaker? The challenge aims to build smarter models for biometrics, security, and real-world AI.

Key Dates:

• Register: Mar 27 – May 10
• Progress: Mar 27 – May 15
• Eval: May 16 – May 23
• Results: May 25
• Paper: June 8

Hosted on CodaBench. Baseline provided. Organizers include Marta Moscati, Saad Saeed, and others.

Tags: Poly-SIM, ACM MM 2026, Speaker ID, Biometrics, AI Challenge, CodaBench, Multimodal Learning