THE VIDEO SHELF · physics and deep learning
Connections between physics and deep learning
Max Tegmark's MIT CBMM talk on why deep learning works so well. He argues that the laws of physics and the symmetries in physical data make it compressible into small neural networks, tying physics to network structure.
Why HN readers recommended it
HN recommends it as a clear argument for why deep and cheap learning succeeds, often alongside the associated paper.
This is a summary from the source dataset, not an endorsement by Hacker News or the video creator.
Video & recommendation details
- Creator / channel
- MITCBMM
- Topic
- physics and deep learning
- Format
- talk
- Level
- intermediate
- HN recommenders in dataset
- 3
- HN mentions in dataset
- 6
- Recommendation years in dataset
- 2016–2026
These counts and years describe the included Hacker News records; they are not YouTube views or the video’s publication date.
From the original HN conversations
Excerpts from the recorded comments. Follow each source for the full comment and surrounding discussion.
Here's a related 2016 talk by Max Tegmark (second author) on connections between deep learning and physics: [VIDEO] The gist of it is that physical data tends to have symmetries, and these symmetries make descriptions of the data very compressible into relatively small neural circuits. …
tedsanders on Hacker News · 2018-05
In Why does deep and cheap learning work so well?You should watch Max Tegmark's talk "Connections between Physics and Deep Learning" if this interests you [1]. Additionally, a paper that has everyone excited about deep connections between the mathematical analysis of physical systems and the hierarchical feature learning paradigm speaks of the connection in …
adityab on Hacker News · 2016-09
In Link Between Deep Neural Networks and the Nature of the UniverseI think the argument for using physics models for constraints in a DL system was a more clear argument [VIDEO]
jonmc12 on Hacker News · 2017-01
In Game Theory Reveals the Future of Deep Learning
Summaries, classifications and recommendation records are from the HackerTube source dataset. Read about this library and its limitations.