Shulou(Shulou.com)11/24 Report--

According to news from CTOnews.com on October 12, ZTE said today that ZTE unveiled a new generation of intelligent computing solution at the China Mobile Global Cooperation Conference held in Guangzhou recently, officially sharing its "latest innovations" in end-to-end computing foundation and large model empowerment with the industry.

According to reports, the scheme includes hardware devices such as AI server, IDC liquid cooling, storage and network, as well as intelligent resource management platform and end-to-end AI development platform. This set of smart solutions support the deployment of Wanka-level large clusters and edge training and push all-in-one machines in the training pool of the center, taking into account the characteristics of leading performance and independent innovation.

Among them, the Wanka-level large cluster in the central training pool supports large-scale parallel computing, reduces training time, and supports complex model training; training and pushing all-in-one machines provide the convenience of rapid delivery and application deployment in edge scenarios. it claims to greatly reduce the cost of private domain large model training and reasoning.

Officials say they will extend the capacity of large models to a wider range of industries through spillover effects. In terms of R & D and application, ZTE Corporation can improve coding efficiency by 30% by using large model technology, and in terms of network operation and maintenance, the reinsurance assistant based on communication model can save 30% manpower investment.

In addition, ZTE Corporation and Shanghai Mobile cooperate to explore anti-fraud applications for upgrading large models. Officials say the prediction accuracy of fraudulent text messages can reach 99%, while the polysemy ability of multi-modal models can be shared among government departments. Urban lifeline applications reduce municipal investment by 15%.

In yesterday's report, CTOnews.com introduced the "smart defense" SMS fraud prevention management system released by ZTE, which is also the industry's first fraud prevention management system based on a large model.

According to reports, under the premise of obtaining customer authorization, the system can effectively improve the recognition accuracy, interception rate and recall rate of fraudulent messages by building an integrated scheme with anti-fraud model as the core, generalized feature neural network and SNS social feature analysis as collaborative, and governance from the root dimensions such as intention and semantics of fraudulent messages.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.