Huawei researchers are focusing on artificial intelligence these days and the latest topic in this subject is Embodied AI (E-AI). The E-AI is one of the most important segments and contributes to new aspects of artificial intelligence technology.
Recently, some Huawei researchers from the Noah’s Ark Lab in Paris released a pre-print research based on the Embodied-AI architecture. The company’s observers believe that giving AI a ‘body’ is the next vital step toward human-level AI tech.
E-AI generally refers to artificial intelligence agents like robots, virtual assistants, and other intelligent systems that can effectively interact and learn in a physical environment. It helps in the formation of human-level AI, also known as strong AI or AGI (artificial general intelligence).
AGI systems can run multiple tasks with the necessary resources. Many tech vendors claim that ‘scale is all you need’ to build AGI systems, which can perform several operations. However, Huawei’s team has a different perspective on this matter.
The pre-print article outlines that large language models like ChatGPT and Google’s Gemini can not understand the real world completely as they do not reside in it. Eventually, the Chinese tech giant says that an AI needs a body to understand actions, memory, and learning. The paper depicts:
“It is a prevalent belief that simply scaling up such models, in terms of data volume and computational power, could lead to AGI. We contest this view. We propose that true understanding is achievable only through E-AI agents that live in the world and learn of it by interacting with it.”
Thus, E-AI will bring the capability of processing raw data from the real world in real time and use it for learning experiences. The system would be able to recognize matters with its own eyes and ears to act as a general intelligence.
Huawei researchers pointed out that there would be thousands of challenges while building up an E-AI technology. Yet, it will lead to better outcomes over the pre-trained AI systems. You can check the entire print HERE.
(Source)
