复旦大学陈涛:不止于VLA,新一代生成式人形机器人运动大模型

Core Insights - The event "Empowering New Energy, Driving the Future" focused on the transformation of achievements by young scientists and the high-quality development of embodied intelligence, gathering over a hundred young scientists and renowned company entrepreneurs [1][3]. Group 1: Technological Innovations - Professor Chen Tao presented a new approach to embodied intelligence, moving beyond the mainstream Visual Language Model (VLA) paradigm, and likened human motion generation to language translation [3]. - The team achieved three core breakthroughs in their motion generation model, enabling precise control of diverse actions and the ability to generate complex body movements solely from natural language instructions [3]. - A novel three-dimensional point cloud multimodal model was developed to address the disconnection between robot actions and their environment, allowing robots to "understand" spatial structures and perform intelligent interactions such as embodied Q&A and path planning [3]. Group 2: Commercialization and Industry Impact - Based on these breakthroughs, Chen Tao's team established a company, Mosheng Intelligent Technology, which has gained attention and recognition in the industry for its pioneering generative Motion series technology [5]. - The exploration by Chen Tao's team signifies a shift in the development of embodied intelligence from theoretical research to practical breakthroughs, contributing to the global embodied intelligence industry [5].