Sama Launches Scalable Training Solution for AI Data Annotation

SAN FRANCISCO, CA / ACCESSWIRE / September 18, 2024 / Sama has unveiled a new flexible, scalable training platform designed to enhance AI data annotation. This proprietary solution aims to reduce project ramp times by 50%, while increasing tag and shape accuracy by 16% and 15%, respectively. As a result, higher-quality models can be produced more swiftly, optimizing time and resources for enterprise clients.

"Basic data annotation training encourages rote memorization," said Duncan Curtis, SVP of AI product and technology at Sama. "Our new platform is more scalable and suitable for complex tasks like automotive and Generative AI data annotation. Annotators can actively learn and receive useful feedback, preparing them for future AI innovations."

This new platform begins with annotation tasks verified by gold answers. Sama's AutoQA™ platform provides autonomous comparison and specific improvement instructions. Annotators can track their progress in real-time, showing significant increases in accuracy and reduced rework. The platform is flexible enough to adapt to changing client needs, ensuring smooth transitions and minimizing rework.

This training solution complements Sama’s suite of products, including some of the largest open-source models. Sama employs a human in the loop (HITL) approach to provide expert feedback, ensuring model performance standards. The company's efforts are backed by SamaAssure™, offering a 98% first batch acceptance rate.

Sama’s investment in this new training platform is part of its broader commitment to responsible AI development and workforce excellence. The initiative also emphasizes Sama’s dedication to promoting diverse perspectives and inclusive talent pipelines.