Academician Dai Qionghai: The Future Understanding and Creation of Artificial Intelligence

Academician Dai Qionghai proposes building a new generation of artificial intelligence starting from brain science, with cognitive computing as the bridge, supported by multimodal circuit observation and multilevel cognitive models. He points out that current AI has limitations in robustness, transferability, energy efficiency, and adaptive capability, and uses the Moravec paradox to explain why robots struggle to understand scenes and object relationships. Tsinghua uses the RUSH-II whole-brain imaging system to observe neuronal connections and has built the PANDA 1-billion-pixel large-scene multi-object data platform.