Discrete Machine Learning
Joscha discusses the challenges of modeling discrete features in generative models and the limitations of deep learning models in representing discontinuities. He proposes the idea of combining deep learning with discrete machines to create more structured and accurate models, highlighting the potential for a more sparse language of thought.In this clip
From this podcast
Machine Learning Street Talk (MLST)
#82 - Dr. JOSCHA BACH - Digital Physics, DL and Consciousness [UNPLUGGED]
Related Questions
Can neural networks be made to reason as discussed in the episode Ilya Sutskever: Deep Learning | Lex Fridman Podcast #94 and the clip Neural Networks and Reasoning?
How does the architecture of neural networks relate to their ability to reason in the episode Ilya Sutskever: Deep Learning | Lex Fridman Podcast #94 and the clip Neural Networks and Reasoning?
How does the architecture of neural networks relate to their ability to reason in the episode Ilya Sutskever: Deep Learning | Lex Fridman Podcast #94 and the clip Neural Networks and Reasoning?