Embodied intelligence (EI), refers to intelligent agents that have physical bodies in the environment and can directly interact with the environment, like humans, they can interact with the environment, perceive, autonomously plan, make decisions, take action, and have the ability to perform a series of smart tasks. As artificial intelligence (AI) systems evolve from basic computing paradigms to comprehensive paradigms covering various perceptual, behavioral, and cognitive patterns, the development of multimodal embodied intelligence computing that is closer to human beings will be a future research trend. Facing challenges such as the scarcity and diversity of multimodal data, the complexity and scalability of embodied intelligence models, the normativity and uniformity of validation environments, and the generalizability from virtual to reality, this forum will invite four experts to summarize the cutting-edge progress in the field of embodied intelligence, analyze and look ahead at current challenges, research frameworks, and future directions, hoping to summarize more consensus, provide more inspiration, and promote the development of the field.
具身智能是指在环境中具有实体、能直接与环境进行物理交互的智能体，像人一样能和环境交互感知，自主规划、决策、行动，并具备执行一系列智能任务能力。随着人工智能系统从基本计算范式向涵盖多种感知、行为和认知模式的综合范式发展，更接近人类的多模态具身智能计算将是未来研究趋势。面向多模态数据的稀缺性和多样性、具身智能模型的复杂性和可扩展性、验证环境的规范性和统一性、以及虚拟到现实的泛化性等方向面临挑战，本论坛将邀请四位专家总结具身智能领域的前沿进展，分析和展望当前挑战、研究框架和未来路线，期望能够总结更多共识、提供更多启迪、推动领域发展。