NVIDIA reports lower Saudi Arabic speech-recognition errors using SDAIA dataset
NVIDIA says fine-tuning Nemotron 3.5 ASR with Najdi and Hijazi speech data reduced word-error rate on its target test split.
Independent intelligence for Gulf AI
Capital, policy, infrastructure & research
Archive
Find GCC AI coverage by company, country, sector or subject.
6 results for “SDAIA”
NVIDIA says fine-tuning Nemotron 3.5 ASR with Najdi and Hijazi speech data reduced word-error rate on its target test split.
Crown Prince Mohammed bin Salman and SDAIA launch the national Year of AI alongside the world's largest sovereign government computing hub in Riyadh.
The national energy titan integrates 7 Terabytes of subsurface data into a proprietary foundation model driving reservoir simulation and predictive maintenance.
The national human capital program hits a historic milestone, bridging the gap between sovereign infrastructure builds and national workforce technical readiness.
The sovereign model powers automated legislative drafting, citizen query routing, and internal administrative search across state agencies.
SDAIA’s membership brings Saudi Arabia into a multilateral forum focused on responsible AI, international cooperation and shared policy practice.