NetEase accelerates the application of large AI models in the industry and will cover more than 100 application scenarios

At the 2023 World Artificial Intelligence Conference (WAIC), NetEase's multiple large-scale artificial intelligence model products were unveiled for the first time, focusing on demonstrating leading innovations in the artificial intelligence infrastructure layer, engine layer, platform layer, model layer, and application layer. Results.

At the booth, the reporter learned that the unmanned loading robot was jointly developed by NetEase Fuxi and China Construction Eighth Engineering Group Co., Ltd. It is a highly intelligent and fully automated main engineering machinery based on the mixing station scenario. On the one hand, it uses unmanned Driving technology enables automated functions such as environmental perception, path planning, active obstacle avoidance, autonomous operation, automatic monitoring and reminders; on the other hand, digital generation technology is used to build an intelligent system for the mixing station, which can utilize comprehensive data from the big data platform Analysis, autonomously select the best to distribute tasks to loaders in the station, and realize intelligent planning and execution of loader operation paths.

NetEase Fuxi Engineering Machinery Technical Director introduced that the unmanned loading robot has technological breakthroughs in four aspects: First, the high-precision positioning and navigation system, the autonomous positioning accuracy is better than 5cm, and the comprehensive automatic driving control accuracy is better than 15cm. Automated operation provides stable and safe support; secondly, the motion control scheme based on learning can quickly adapt to different models, and can automatically perform actions such as movement, acceleration and deceleration, shoveling and unloading, and the bucket full rate of shoveling operations reaches more than 95%, and the unloading rate reaches more than 95%. The residual rate of the material hopper is less than 5%, and the shoveling efficiency is close to that of the old master. Third, a high-precision visual perception algorithm is used to realize real-time detection of materials in the silo and hopper, and dynamically analyze the production situation in the mixing station to monitor errors in real time. Less than 5%; fourth, high-precision scene reconstruction and rendering technology are used to realistically restore the 3D scene of on-site operations. Through the simulation and prediction functions of digital twin scenes, holographic control of the operation process is achieved, improving production efficiency and work safety.

At the same time, the reporter learned that during this World Artificial Intelligence Conference, NetEase Fuxi also jointly launched the first AI painting experience event with the Shanghai Municipal Youth League Committee and Donghao Lansheng. The event covered the entire exhibition area of ​​the World Artificial Intelligence Conference, and participants Just enter the keywords of the work and select different painting styles to generate exquisite paintings. This AI painting experience event is technically supported by NetEase Fuxi-Danqing Yue, and relies on NetEase Fuxi's self-developed image generation model - "Danqing".

Danqing is a Chinese model self-developed by NetEase Fuxi. It is trained based on native Chinese corpus data and NetEase's own high-quality image data. It is a 100% domestic large-scale model. The model training data has undergone strict text and image review. While ensuring that the data source is compliant and the generated content is compliant, the model has better Chinese understanding capabilities and the works created can better meet Chinese aesthetics.

Prior to this, NetEase CEO Ding Lei stated at the Zhejiang Provincial Platform Economy High-Quality Development Conference that exploring the innovation and application of AI large models as quickly as possible is NetEase’s top priority.

It is understood that NetEase AI large models are accelerating to cover more than a hundred industrial application scenarios. While accelerating innovative applications, NetEase Fuxi Youling crowdsourcing platform will also provide 100,000 people with new AI occupations this year, including remote excavator Human-machine collaboration jobs such as drivers, AI painters, and AI expression binding artists. We will also provide job seekers with personalized training based on their user profiles to help them improve their abilities and increase their income.

The above is the detailed content of NetEase accelerates the application of large AI models in the industry and will cover more than 100 application scenarios. For more information, please follow other related articles on the PHP Chinese website!