Foreign media: Chinese scientists have made the first discovery that large language models (LLMs) can spontaneously form concept systems similar to those of humans, enabling the classification and understanding of natural objects, which is an important foundation of human cognition. This discovery provides new evidence for the debate on whether artificial intelligence has human-like cognitive abilities, indicating that AI systems may simulate key features of human thinking.
The research was published in the journal Nature Machine Intelligence, pointing out that LLMs and multimodal large models (MLLMs) can establish object representations close to those of humans by processing data such as language, images, and audio.
This advancement is of great significance for gaining a deeper understanding of the potential of AI thinking and promoting research into human-like intelligence, also demonstrating the enormous potential of AI in the fields of cognitive science and the simulation of human intelligence in the future.
Original article: https://www.toutiao.com/article/1834894212413508/
Disclaimer: The article represents the views of the author only.