DeepMind公司人工智能研发组织模式分析

Analysis of DeepMind’s Artificial Intelligence R&D organization mode

摘要:

国际上大型科技公司正在以收购公司、设立分支机构等方式加快建设人工智能实验室，在人工智能研发方面抢占先机。人工智能实验室DeepMind被谷歌母公司Alphabet收购以来，以实现通用人工智能为核心目标，构建矩阵型、网络开放式的研发组织结构，通过灵活的项目运作模式、跨学科的工作方式、长期的资本投入和高水平的人才培养体系，在生物、医疗、能源等多个领域取得重大突破。其研发组织模式的做法和经验对中国加强人工智能前瞻性研发部署、企业设立人工智能实验室、产学研用深度融合和开放合作等方面提供一定的启发和借鉴。

关键词:人工智能;人工智能实验室;研发组织模式;跨学科;矩阵结构

Abstract:

Large technology companies are accelerating the construction of artificial intelligence laboratories by acquiring companies and setting up branches to seize opportunities in artificial intelligence (AI) R&D. Since its acquisition by Google’s parent company Alphabet, DeepMind has built a matrix-type, open-network R&D organizational structure with the core goal of achieving general AI. Through flexible project operation mode, interdisciplinary working methods, long-term capital investment and high-level talent training system, DeepMind has made a major AI breakthrough in many fields such as biology, medical care, and energy. It’s unique R&D organization mode provide certain inspiration and experience for China to strengthen the deployment of forward-looking R&D of AI, encourage enterprises to set up AI laboratories, deepen in-depth integration of industry, academia, research and application, and strengthen open cooperation in AI.

Key words:artificial intelligence; artificial intelligence laboratory; R&D organizational mode; interdisciplinary; matrix structure