An industrial consortium led by GTD will develop strategies to accelerate the adoption of AI by industry, based on the concept of AI 'capsules'. This 4-year project is led by GTD SCIENCE INFRASTRUCTURES & ROBOTICS, with the participation of PAL Robotics, Solver Machine Learning and ZYLK.net, supported by the technology centers BSC-CNS Centro Nacional de Supercomputación and IKERLAN. The goal of the project is to universalize the use of AI in industry by establishing a simpler access threshold that can simplify the adoption of pre-engineered, integrated solutions for a wide variety of applications. CAPSUL-IA proposes the concept of AI model containers tailored to different problems, and with multiple trade-offs between cost, accuracy, speed of inference, and robustness - among other metrics.
