New Framework For ASR Compute Scaling
A new framework allows for test-time compute scaling in Automatic Speech Recognition (ASR) systems using depth-conditioned looped transformers. End-to-end ASR systems typically use fixed-depth acoustic encoders at inference.
Topics
Developing
- 884d Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore.
- 884d Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur.
- 884d Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est.
- 884d Sed ut perspiciatis unde omnis iste natus error sit voluptatem accusantium doloremque laudantium.
Sources · 7 independent
Modernity/arxiv
“Test-Time Compute Scaling for ASR with Depth-Conditioned Looped Transformers. Authors: Yacouba Kaloga, Shashi Kumar, Shakeel A. Sheikh, Driss Khalil, Petr Motlicek, Ina Kodrasi Abstract: End-to-end ASR systems typically use fixed-depth acoustic encoders at inference, making it...”
Unlock the full story
Get a Pro subscription or above to see the live story progression and the full list of independent sources confirming each event as they happen.
Log in to upgrade