SEOUL -- South Korea's Hyundai auto group has partnered with a state information agency to share training data for artificial intelligence development. The data will be used for upgrading the group's future mobility technologies such as voice recognition, autonomous driving and mobility-related software.
A training dataset is a dataset of examples used during AI model development's machine learning phase. The dataset includes a set of examples used by an AI neural network to study and construct algorithms. Training datasets are basically the food for the AI system's brain. The more the dataset, the smarter AI gets.
Hyundai said in a statement on April 8 that the group signed a memorandum of understanding with the National Information Society Agency (NIA) to share AI training datasets. NIA will provide datasets that consist of information about natural language, on-road structures, streets, street lanes and video content of human movements.
The group will use datasets including free talking conversation and slang to upgrade its voice recognition technology. Vehicle operation-related data will be used for the upgrade of autonomous driving technologies. Robotics and infotainment technologies will also be upgraded.
"Through this MOU, we will accelerate the transition into a smart mobility solution group and offer our customers new mobility services by upgrading AI technologies," Hyundai auto group's vice president Kim Dong-wook was quoted as saying.
Hyundai has gone full throttle to develop and upgrade future mobility technologies to check American and European rivals. In 2020, the group revealed a plan to connect all models produced from 2020 with an integrated infotainment platform. Some $25 million was invested in Envisics, a British dynamic holography and auto head-up-display (HUD) developer, to develop a next-generation augmented reality HUD that will increase the safety and convenience of drivers.
© Aju Business Daily & www.ajunews.com Copyright: All materials on this site may not be reproduced, distributed, transmitted, displayed, published or broadcast without the authorization from the Aju News Corporation.