hackertube.io

Connections between physics and deep learning

MITCBMM · talk · Watch on YouTube

Machine learning & deep learningFundamental physics

What it is

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 Hacker News recommends it

HN recommends it as a clear argument for why deep and cheap learning succeeds, often alongside the associated paper.

  • 3 people recommended it
  • 3 different months
  • 1 strong recommendations
  • first 2016, last 2026

What HN commenters said

  1. “physical data tends to have symmetries, and these symmetries make descriptions of the data very compressible into relatively small neural circuits”

    Physical data has symmetries, making it compressible into small neural circuits.

    tedsanders, 2018-05 · Strong recommendation · in “Why does deep and cheap learning work so well?”

  2. “You should watch Max Tegmark's talk "Connections between Physics and Deep Learning" if this interests you [1].”

    Tegmark's talk linking physics and deep learning.

    adityab, 2016-09 · Recommendation · in “Link Between Deep Neural Networks and the Nature of the Universe”

  3. “the argument for using physics models for constraints in a DL system was a more clear argument”

    Clear argument for physics-based constraints in deep learning.

    jonmc12, 2017-01 · Recommendation · in “Game Theory Reveals the Future of Deep Learning”

Excerpts from Hacker News comments, linked to the originals. Summaries and labels were written by a language model from those comments.

Topics: deep learning, physics, tegmark, neural networks, symmetry

Also recommended by the same people