Project Overview

To raise patent-examination quality, reduce examiners’ workload, and provide more convenient, faster services to the public, the existing system is being upgraded for intelligence, so that patent examination, patent search, and public services move toward intelligence and innovation. The intelligent upgrade will greatly strengthen information-service capability, help raise examination quality and efficiency, strengthen data-resource management, support continuous improvement of administrative efficiency, optimize the information-infrastructure environment, strengthen information security and O&M services, and help raise informatization capability for international cooperation.

The intelligent patent-examination system uses a “platform + modules” construction model spanning the full process of application, acceptance, examination, reexamination, and invalidation. It applies artificial intelligence, big data, cloud computing, microservices, and other modern information technologies to realize core functions such as intelligent search, online translation, graphic recognition, application-quality monitoring, and fine-grained examination-cycle management. It emphasizes overall integrity, compatibility, and stability, and enhances scalability and flexibility, fundamentally raising IP examination quality and efficiency across the board so that the public has a greater sense of gain.

Solution Architecture

The intelligent patent-examination system uses “platform + modules” to resource infrastructure and servitize applications. A “high cohesion, low coupling” microservice framework raises business flexibility and maintainability and better supports the business. It mainly includes a basic cloud platform, a big-data platform, business modules, business applications, a security system, and overall O&M.

The basic platform deploys Sugon AI servers in a CPU+GPU technical architecture, combined with Alibaba proprietary cloud, to provide integrated AI computing resources and services.

Value and Benefits

1. AI deep-learning development makes examination and search more intelligent and convenient

The Sugon AI server is a high-performance dual-socket server mainly for deep-learning users. It supports up to 3 TB of memory, integrates a high-performance SATA controller, supports up to 24 hot-swap 2.5-inch drives, and supports 8 dual-width or 16 single-width NVIDIA GPU accelerator cards. This expansion capability is sufficient for heavy workloads and suits deep-learning users’ offline network-training environments.

For CNIPA’s intelligent patent examination and search platform, Sugon uses a “high cohesion, low coupling” microservice framework and “platform + modules” to resource infrastructure and servitize applications. Sugon AI servers are applied to the basic cloud platform, big-data platform, and business modules, using the strongest computing power to raise patent-search quality and efficiency and support high-quality development of examination business.

2. Artificial intelligence drives technological innovation and injects new vitality into patent applications

Facing rapid development of AI technology, customer application innovation will face greater opportunities and challenges. Sugon provides deep-learning solutions with unique strengths, supporting patent-application systems with its own software platform and high-quality hardware. Artificial intelligence drives technological innovation and injects new vitality into patent applications.