Fudan's Jia Xiaosong says model architecture is not the bottleneck in embodied AI
Overview
Fudan University assistant professor Jia Xiaosong says the real bottleneck in embodied AI is not model architecture, according to Leiphone (雷峰网).
He says VLA and WAM architectures keep emerging, and he draws on his work moving from trajectory prediction to end-to-end autonomous driving, reinforcement learning and world models. The report does not say which bottleneck he identifies instead.
Written by AI from the articles below · updated Oct 19, 1:33 AM ET
Check the sources:
Article timeline
The articles in this story. Times are ET.
- LeiphoneNewsFudan's Jia Xiaosong says VLA and WAM architectures keep emerging, but model architecture is not the real bottleneck
AIFudan University assistant professor Jia Xiaosong says the real bottleneck in embodied AI is not model architecture, drawing on his experience moving from trajectory prediction to end-to-end autonomous driving, reinforcement learning, and world models.
Heat Trend
Not enough continuous observations to show a trend yet.