news

한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina

the large language model (llm), which will start to perform amazingly in 2022, has not only set off a "battle of hundreds of models" in its own field, but also made the humanoid robot track become more popular in the past two years.

"humanoid robots are not a new topic. we believe that embodied intelligence will be the last piece of the humanoid robot technology puzzle, and this is just the beginning." at the transform·2024 global digital ecology conference held during the 3rd global digital trade expo, jiang lei, chief scientist of humanoid robot (shanghai) co., ltd. said this.

similarly, embodied intelligence is not a new concept, but it has not progressed much in the past few decades. the emergence of large language models has brought new possibilities to the design of intelligent agents, one of the important elements of embodied intelligence.

what is the impact of llm on embodied intelligence? let’s go back to a basic question: should we make a “human-like” humanoid robot? does china have the potential to emerge as a “giant” in the field of embodied intelligence? at the above-mentioned conference, industry professionals, including jiang lei, held a dialogue and collision around these issues.

jiang lei said that for robots, “we prefer to communicate with robots in a more natural way rather than through complex code programming.” because of this, the preferred method of almost all humanoid robot companies now is to use big language model-driven, "this thing has already worked out, that is, the robot's planning, decision-making, and interaction with humans can be closed-looped using large language models."