entering the deep waters of the "city of artificial intelligence", china's first "car-road-cloud" data collaboration platform was released
2024-09-26
한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina
cover news reporter fu wenchao
on september 24, under the guidance of beijing high-level autonomous driving demonstration zone working office, kaiwang data and beijing international computing service co., ltd. jointly organized the "digital intelligence acceleration: data power and application innovation under ai big model" ecological conference in yizhuang, beijing. nearly 200 smart driving and artificial intelligence experts and representatives discussed the strategic value of data in the context of ai big models.
it is reported that beijing yizhuang is accelerating the construction of an "artificial intelligence city" in all areas. on the same day, kaiwang data, together with jingsuan company and chewang company, wholly-owned subsidiaries of yizhuang intelligent institute, built the country's first "vehicle-road-cloud data collaboration platform" based on the "artificial intelligence data training base", and it has been officially launched. the platform focuses on autonomous driving vehicle-side perception data and vehicle-road-cloud intelligent transportation scenarios, providing efficient and secure intelligent data application services for demanders, not only providing accurately adapted and processed data, but also providing gpu computing power cluster services required for model training. the first batch of project transaction signing partners include ecological enterprises such as guoqi intelligent control and huixi intelligent.
"data, algorithms and computing power are the three core elements that constitute artificial intelligence technology." ma guang, assistant to the general manager of beijing international computing power service co., ltd., said that from the perspective of external conditions, the most obvious problems currently restricting the development of large-scale models of enterprises are computing power and data. therefore, building a public computing power platform will solve the problem of insufficient ability of enterprises, especially entrepreneurial and small-scale enterprises, to purchase computing power during development.
ma guang mentioned that the artificial intelligence data training base is not only a platform for data aggregation, cleaning and labeling, but also a core carrier for building a safe and compliant data circulation model and a trusted infrastructure for the safe circulation of high-quality data. it provides full-process services for trusted data storage, trusted processing, and trusted delivery, as well as basic capabilities such as infrastructure, application platforms, and external portals. it brings together tool capabilities such as data labeling, model training, and compliance assessment of ecological partners, and provides an efficient and secure path for data circulation by introducing a sandbox supervision mechanism.
around the country's first "vehicle-road-cloud data collaborative platform", beijing computing company, kaiwang data, guoqi intelligent control, huixi intelligent and other parties have signed relevant strategic cooperation agreements, clarifying that they will cooperate in their respective areas of need and advantage in the future, focusing on "big models + big data + big computing power", supporting the construction of beijing's artificial intelligence data training base, and is expected to further strengthen the full-factor ai native industry foundation of "high-quality data sets - intelligent computing power base - autonomous and controllable algorithms - demonstrative application scenarios", and maximize data utilization efficiency and realize data value.
"based on the industry's existing data, we can achieve industry data sharing and data reuse through platform storage, supervision and reprocessing. the industry efficiency solves the company's doubling cost reduction needs and can help companies achieve the data application advantage of 'half the cost and three times the speed'. for example, with the convergence of sensor models, locations, and algorithms, we have the opportunity to match similar ones, and for those that don't match, we can achieve personalized results through fine-tuning and secondary processing." yu xu, founder and ceo of kaiwang data, said that with the launch of the platform, data sharing in the field of intelligent transportation in the future will greatly reduce data circulation costs and greatly improve data utilization efficiency. the "vehicle-road-cloud data collaboration platform" is an innovative model that combines the concept of artificial intelligence data training bases to explore data services and new data circulation forms in the field of intelligent transportation, which maximizes data utilization efficiency.
