Episode 07: Yujia Huang, Caltech, on neuro-inspired generative models
Generally Intelligent
00:00
The Future of Reasoning in Deep Learning
I would be excited to see that the neural networks can react more dynamically towards different goals, maybe inform mathematical or for vision. In general, I hope we can have these new components into deep learning. This is a replacement of the current forward only neural networks. You can replace the things with these sort of dynamics with feedbacks. That could be caught at reasoning but we just need to provide a better definition of reasoning.
Transcript
Play full episode
