Document Download

Pseudo-Labels as Rewards, Not Targets: Group-Relative Policy Optimization for Semi-Supervised ASR

Author / Organization: Shashi Kumar, Yanis Labrak, Sergio Burdisso, Esaú Villatoro-Tello, Petr Motlicek, Kadri Hacioglu, Andreas Stolcke

File Format PDF (Portable Document Format)
Publication Year 2026
Source Host arXiv (arxiv.org)
License / Terms arXiv Open Access / Creative Commons

Ready for Immediate Access

Your selected document is ready for instant access with zero mandatory waiting period.

Alternative Options: Download via MainReads Proxy Read Online in PDF.js Viewer Open Source Website »
Linking & Access Policy: MainReads provides direct outbound links to documents and educational resources hosted by original source repositories, universities, and open libraries. Downloads are never made conditional on advertisements or user registrations. If you encounter an inaccessible or broken resource link, please use the Report Broken Link feature on the catalog page.
« Return to Document Detail & Online Reader