In this video upload available on Spotify (we'll try this once and see how it's received), we revisit Reinforcement Learning (from way back in episode 28) and this time discuss how to turn it into Deep Reinforcement Learning by swapping out the Q-Table and putting a neural network in its place. The end result is a sort of 'bootstrapping intelligence' where you let the neural net train itself. 
We also discuss: 

 How this, if at all, relates to animal intelligence. 
 Is RL a general purposes learner? 
 Is it a path to AGI?

Links:

 Github Code Base
 Presentation Slide Pack
 Youtube version


--- 

Send in a voice message: https://podcasters.spotify.com/pod/show/four-strands/message
Support this podcast: https://podcasters.spotify.com/pod/show/four-strands/support

Episode 43: Deep Reinforcement Learning

A podcast with episodes loosely tied together by Popper-Deutsch Theory of Knowledge. David Deutsch's 4 Strands tie everything together, so we discuss everything we find interesting be it science, philosophy, computation, politics, or art. But there is a heavy emphasis on the exploration of intelligence and the search for Artificial General Intelligence (AGI).
 Support this podcast: https://podcasters.spotify.com/pod/show/four-strands/support

Society & Culture

Personal Journals

History

Philosophy

Places & Travel

How do humans form 'fuzzy categories'? How does this all relate to essentialism? Is essentialism false? Or is it partially true? And how does this all relate to Critical Rationalism?

Picking up where we left off last week, Bruce gets
deeper into Douglas Hofstadter’s ideas on language and the mind and his assertion that “analogy-making lies at the heart of intelligence.”

Bruce considers how Hofstadter’s theories may be interwoven with ideas on language and cognition promoted by Steven Pinker in "How the Mind Works" along with, as usual, the epistemology of Karl Popper and David Deutsch.

We again consider if this is an inductive theory? And how should critical rationalists view theories like this?

Follow us on Twitter: ⁠https://twitter.com/bnielson01⁠

--- 

Send in a voice message: https://podcasters.spotify.com/pod/show/four-strands/message
Support this podcast: https://podcasters.spotify.com/pod/show/four-strands/support

Episode 86: Fuzzy Categories, Essentialism, and Epistemology (Hofstadter Part 2)

This is the first of our two part series (that may or may not be released back-to-back) where Bruce delves into the work Douglas Hofstadter, specifically the book Surfaces and Essences. We consider what is the relationship—if there is any—between critical rationalism and Hofstadter's idea that analogy is a core mechanism of human cognition. Is it fair to criticize Hofstadter's ideas as being inductivism in disguise? Could something like what Hofstadter suggests (i.e. analogy) be central to human consciousness and creation of AGI?

Follow us on Twitter: ⁠https://twitter.com/bnielson01⁠

--- 

Send in a voice message: https://podcasters.spotify.com/pod/show/four-strands/message
Support this podcast: https://podcasters.spotify.com/pod/show/four-strands/support

Episode 85: Critical Rationalism and Douglas Hofstadter (Part 1)

Here we discuss a 1992 interview with David Deutsch where he makes the case that video games are inherently educational, not addictive, and that children should not be stopped from playing as much as they want. We contrast the view of humans, science, and knowledge promoted there by David Deutsch with the more pessimistic view of thinkers such as Jonathan Haidt today. Bruce and Peter reflect on their own mixed feelings on this issue both as critical rationalists and parents. 

David Deutsch on video games: 
https://takingchildrenseriously.com/video-games-a-unique-educational-environment/
Peter briefly quotes from this recent article by Jonathan Haidt:
https://www.thefp.com/p/jonathan-haidt-worried-about-the-boys-too

Follow us on Twitter: https://twitter.com/bnielson01

--- 

Send in a voice message: https://podcasters.spotify.com/pod/show/four-strands/message
Support this podcast: https://podcasters.spotify.com/pod/show/four-strands/support

Episode 84: Are Video Games Harmful to Children?

Bruce summarizes his (unique?) understanding of Karl Popper’s epistemology that (possibly?) straddles the line between orthodox and unorthodox and is Influenced both by Deutsch, more old school Popperians, and his own unique interpretation of critical rationalism. 

Bruce claims that the key difference between regular "folk epistemology" (i.e. how humans reason without a correct understanding of epistemology) and "Popper's epistemology" (aka "Critical Rationalism" or the correct epistemology) is due to Popper's epistemology having a 'second axis' that regular folk epistemology entirely lacks. This 'second axis' is rooted in a choice to make your theories bold and risky by maximizing empirical content.


This makes Popper's epistemology 2-dimensional instead of 1-dimensional. 


If this fact is missed, Bruce claims your epistemology collapses back to be regular old folk epistemology and you are no longer doing critical rationalism.

Refutation, corroboration, explanation, induction, falsification, verisimilitude, “the Popperian war on words,” and “Popper’s ratchet” -- from past podcasts! -- are all touched upon.

Follow us on Twitter: ⁠https://twitter.com/bnielson01⁠

--- 

Send in a voice message: https://podcasters.spotify.com/pod/show/four-strands/message
Support this podcast: https://podcasters.spotify.com/pod/show/four-strands/support

Episode 83: Popper's Second Axis (aka Bruce's Epistemology?)

In an episode that may (or may not) be his magnum opus, Bruce introduces his term for Karl Popper’s idea that you are only allowed to solve problems with your (scientific) theory by making it more empirical, not less empirical. 
Bruce makes the case that this is one of Karl Popper’s least appreciated ideas, as all of us are tempted by ad hoc saves that move our ideas in the direction of vagueness. 
Bruce also considers where conjectures come from and if Popper thought there existed a scientific method.

--- 

Send in a voice message: https://podcasters.spotify.com/pod/show/four-strands/message
Support this podcast: https://podcasters.spotify.com/pod/show/four-strands/support

Episode 43: Deep Reinforcement Learning

Download our free app to listen on your phone