In the era of large models, intelligent computing infrastructure serves as the underlying driver for original model innovation and application enablement, and forms the foundational base for artificial intelligence to empower new industrialization. This paper analyzes and elaborates on the development trends and challenges of intelligent computing infrastructure. From aspects such as hardware-software coordination, computing power efficiency, network architecture, and operation and maintenance support, it summarizes four key points for evaluating the capabilities of intelligent computing infrastructure for large models, and proposes strategic suggestions for developing the ecosystem of intelligent computing infrastructure.
China’s new industrialization has entered a new historical stage and is heading towards the high-quality development when artificial intelligence serves as an engine. As one of the key elements driving the development of artificial intelligence, data are rapidly integrating into various aspects such as production, distribution, circulation, consumption, and social service management, becoming the cornerstone for achieving intelligence, automation, and optimizing industrial processes. This article analyzes the current dilemma of traditional and emerging risks of artificial intelligence data in the industrial field, such as “insufficient to use”, “afraid to use”, “difficult to use”, and “unwilling to use” from a systematic research perspective. It points out that the basic logic of governance is to balance development and governance, and reveals governance strategies such as following the “guidance” of legal rules and adopting the “technique” of self-discipline.
This paper systematically reviews and analyzes the significant innovation directions based on the Transformer architecture. It examines the evolution of large language model architecture from three dimensions: innovation within the Transformer architecture itself, fusion innovation with other architectures, and innovations in non-Transformer architecture. This paper also provides an outlook on the future development directions of foundation models.
Artificial intelligence is becoming the core driver of new industrialization, leading society into the era of intelligent economy. The large model technology demonstrates powerful capabilities in multiple tasks through pre-training and fine-tuning. However, its application faces challenges in terms of computational power, data, algorithms, and security. The large model application and development platform (MaaS) enhances the flexibility and efficiency of enterprises in utilizing large model technology through standardized services. This paper provides the key technical pathways for constructing the Xingchen MaaS platform, offering a reference for enterprises in their intelligent development.
With the rapid development and extensive application of Artificial Intelligence (AI) technology, enterprises are actively introducing AI technology to achieve intelligent upgrades and transformations of their businesses, so as to enhance competitiveness. This paper proposes a new operator IT support model based on the three principles of decoupling, convergence, and interconnection, namely AI+ MaaS. This paper describes the exploration and practice of this model in platform architecture, model orchestration, business interconnection, unified computing power operation, and precise intent recognition. This approach addresses the challenges of traditional IT architectures in terms of flexibility, compatibility, and scalability to drive scalable value operations for AI+ businesses.
Firstly, the application progress of Neural Radiance Field (NeRF) technology in communication industry is discussed. Secondly, the basic principle of NeRF and the construction method of communication digital twin are introduced. Finally, the application of NeRF in 3D modeling of communication facilities, network planning is introduced, which aims to provide valuable reference for researchers in the communication industry and stimulate future research directions.
In the critical period of building modernization and promoting new industrialization, the rise of Artificial Intelligence(AI) technology has provided a strong impetus for the transformation and upgrading of the manufacturing industry. The landing of AI technology in the whole process of manufacturing has become an important engine to promote the industrial upgrading of the manufacturing industry. Firstly, the significance of AI technology enabling the manufacturing industry is analyzed. Starting from the core links of the manufacturing industry such as research and development design, production and manufacturing, operation management, and product services, the key technical points of the integration and application of AI technology in the entire process are elaborated in depth. Secondly, the application of AI technology in the manufacturing industry is introduced, and the challenges of applying AI technology throughout the entire manufacturing process are deeply analyzed. Finally, the technological development trend of AI in manufacturing industry is forecasted.
This paper explores the current application of Artificial Intelligence (AI) in industry. It analyzes the policy orientations of some major economies, highlighting the role of AI in enhancing industrial innovation capabilities, optimizing industrial structures, strengthening supply chain resilience, and promoting green industrial development. As AI is an element of the new quality productive forces and digital economy, its integration with industry is critical for advancing industrialization. Through the fusion of large and small models, edge intelligence, and cognitive manufacturing, this study reveals the pivotal role of AI in driving the higher-level intelligent transformation of industry. The findings indicate that AI not only improves industrial efficiency, but also facilitates structural optimization, providing strong support for achieving in-depth empowerment of industrialization.
With the advancement and innovation of technology, automobiles are evolving from traditional mobile transportation tools to intelligent mobile spaces and even intelligent beings. Intelligent cockpit is the main entrance to human-machine interaction. Beyond integrating functionality and entertainment, the development of Artificial Intelligence (AI), particularly general AI, poses numerous requirements for the cockpit, such as autonomy, adaptability, and personalization. To meet these demands, this paper designs a intelligent cockpit system based on large language model and agent, and outlines the technical implementation path of this system through a cloud-based architecture. Starting from application scenarios within the cockpit, this paper constructs a highly intelligent cockpit system with autonomy, adaptability, and personalization.
With the continuous development of space technology, space control software is facing increasingly complex business forms and continuously changing business needs. Software shows the trend of increasingly complex functions, larger scale, shorter development cycle, etc. Software systems become increasingly large and difficult to control. When engineers face new software requirements, they tend to find the most similar function points in the historical model tasks for software code inheritance, so as to reduce development costs and improve development efficiency. In the face of massive code base, it is difficult for code retrieval methods to obtain the requirements and codes related to intent efficiently and accurately, which affects the efficiency and quality of software development and restricts the efficient and reliable delivery of software. In order to solve the above problems, this paper proposes an artificial intelligence(AI)-based code retrieval method for space control software. Based on semantic vector model, this method extracts and processes documents and codes in historical assets of type control software, constructs a mapping relationship between function codes and requirements, and constructs a retrieval library to provide support for the code. This paper focuses on the practical application of AI enabled industrial software design to optimize software development process, improve development efficiency and quality, and enhance intelligent capability. It provides effective solutions to solve the problem of efficient utilization of existing assets in many research and development scenarios in the industrial industry, and provides strong support for promoting the high-quality development of intelligent manufacturing in China.
A data auditing and security management system based on user and entity behavior analysis offers an innovative solution to address internal and external data security threats in enterprises. By continuously monitoring the behaviors of users, devices, and applications, the system establishes dynamic baselines and performs real-time anomaly detection to identify potential threats effectively. This paper proposes a framework centered on four key elements—entities, behaviors, baselines, and algorithms—along with a three-step process comprising data collection, behavior analysis, as well as response and handling. It demonstrates how artificial intelligence-enhanced user and entity behavior analysis (UEBA) can be used to construct an intelligent data security auditing system, strengthening data protection capabilities and ensuring compliance.
This paper refers to satellite landscape data and urban planning data published by the government, and uses large model technology to expand the sample size. Then, the sample data are used for multi-model joint learning training. Through model training, the high-definition satellite base map and multi-spectral data of Beijing urban area are measured and restored with high precision, including the urban building base, vegetation, and road. The large language model technology is used to program three-dimensional modeling, and finally automatically generates in the Unreal Engine. The process reproduces landscape assets within a 1 000 square kilometer area of central Beijing. This method not only improves the efficiency and accuracy of urban modeling, but also provides a new research tool and perspective for urban planning, historical preservation and other fields.
In the context when global resource constraints intensify and China makes efforts to build a circular economy development model, this paper explores the main technical means of artificial intelligence technology applied to the establishment of circular economy based on the basic model of circular economy development, and builds a framework of the application model of artificial intelligence empowering circular economy establishment. Meanwhile, this paper analyzes the main challenges faced by artificial intelligence technology empowering circular economy establishment at the current stage, and puts forward corresponding development suggestions.