Mark Hasegawa-Johnson, Xiuwen Zheng, Subhashini Venugopalan, Dhruuv Agarwal, Venkatesh Ravichandran, Colin Lea, and Edward Cutrell, "Speech Accessibility Project Challenge 2: Advancing Universal Dysarthric Speech Recognition from Offline to Real Time"
Abstract
SAPC2 builds on the success of the Interspeech 2025 Speech Accessibility Project Challenge, which demonstrated significant progress in dysarthric speech recognition — reducing Word Error Rate (WER) from the Whisper-large-v2 baseline of 17.82% to 8.11%. This new edition introduces a larger, more diverse, and etiology-balanced corpus, further promoting fairness, robustness, and inclusivity in impaired-speech ASR. The challenge invites the research community to push the state of the art, develop innovative modeling techniques, and set new standards for accessible speech technology. The challenge featured two complementary tracks: (1) an unconstrained ASR track, in which participants were permitted to use models of any size or architecture, aiming to advance the state of the art in dysarthric speech recognition, and (2) an efficiency-constrained ASR track, in which submitted systems had to meet strict limits on model size and inference time, promoting lightweight and deployable solutions for real-world use.