OrukLabs Cuts Speech Recognition WER from 4.936% to 4.757% with RL Recipe, Open-Sources Weights
DrDatta_AIIMS · x · 2026-11-12
OrukLabs says it has cracked RL for speech recognition by stopping positive credit assignment on wrong words. The full recipe cut mean WER from 4.936% to 4.757% across nine public English test sets, and the weights are released openly. A practical, reproducible improvement for anyone building ASR-based products.
More from Research
- 1920 Popular Mechanics clipping shows eclipse that confirmed light bends under gravity — generativist · 2026-11-12
- ELIZA's Lost Code: 2,600 Lines Found in MIT Archives in 2021, Retyped by Hand — demian_ai · 2026-11-12
- 1.5-Year Study: Cultural Bias in AI Curiosity and Its Role in Multi-Agent Collaboration — radamihalcea · 2026-11-12
- ContactBench 1.0: A Benchmark for Contact-Rich Robot Manipulation With Printable Assets — leto__jean · 2026-11-12
- Formalizing all of nLab joked as the 'final boss' of AI math — burny_tech · 2026-11-12
- Community ports Illumina's PromoterAI genomics model to PyTorch with new optimizations — jmschreiber91 · 2026-11-12