Deep Learning Methods Can Be Considered Tools for Scientific Progress

2min Snip

00:00

Play full episode

Summary

Transcript

Episode notes

The relativists notion of causality is ingrained in relativity. There are light cones, but the way that it is used is very much colored by the fact that physicists know there is an arrow of time. Id wich w y c h says, can deep learning methods like deep noural networks and transformers really be considered tools for scientific progress when their underlying models aren't currently interpretable or meaningful?

Welcome to the July 2022 Ask Me Anything episode of Mindscape! These monthly excursions are funded by Patreon supporters (who are also the ones asking the questions). I take the large number of questions asked by Patreons, whittle them down to a more manageable size — based primarily on whether I have anything interesting to say about them, not whether the questions themselves are good — and sometimes group them together if they are about a similar topic.

Big news this week! Mindscape is working with Bold.org to sponsor a college scholarship for students interested in studying the fundamental nature of reality. Listeners can find more details and donate here. Our immediate goal is to raise $10,000, and I will match the first $5,000, so this shouldn’t be too hard for us here. Hopefully we can raise much more! And hopefully this will help encourage someone who might not otherwise have been able to study this kind of topic.

Support Mindscape on Patreon.

See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.