Connexion

← Toutes les vidéos

On Memory as a Self-Adapting Agent

vidéo · Carlos Farias · mise en ligne le 4 novembre 2024

« Your most important collaborator is you 6 months ago, but he's not answering emails. »

Memory Transferred between Snails, Challenging Standard Theory of How the Brain Remembers
- Scientific American

Unlocking nature's code: Researchers draw parallels between AI models and genetic encoding
- Trinity College Dublin

Basal Xenobot transcriptomics reveals changes and novel control modality in cells freed from organismal influence
-  Communications Biology

Paul Davies - Why There is ‘Something’ Rather than ‘Nothing’ ?
- Closer To Truth

We are driven to make stories about ourselves in our world that make sense. 
Because all of us are trying to continuously modify our models of ourselves to make things make sense.
Humans are fundamentally storytellers. This isn't just about humans, this is about all good agents.
If you are not a good storyteller in a primitive way, making actionable models of yourself in your outside world, you will never get out of the single cell phase.
In fact, you will not survive as a single cell. You will not survice as a persistent chemical pathway either.
That fundamental thing of active inference loop that causes you to store some priors and to try to figure out what exactly is going on in a way that is going to allow me to make the decisions which are coming up constantly. You're going to run out of energy, die and be eaten very quickly if you're not constantly taking actions.
That requires modeling and that does not wait till we get to human stage. That was there from day one of evolution and possibly before that.

We discuss Michael Levin's paper "Self-Improvising Memory: A Perspective on Memories as Agential, Dynamically Reinterpreting Cognitive Glue." Levin is a scientist at Tufts University, his lab studies anatomical and behavioral decision-making across biological, artificial, and hybrid systems. His work spans developmental biology, artificial life, bioengineering, synthetic morphology, and cognitive science.

Next, watch my first interview with Michael Levin...
What are Cognitive Light Cones? https://youtu.be/YnObwxJZpZc?si=qRksKq3xcmbtcRAW

1. Memories as Agents
0:00 Introduction
1:40 2024 Highlights from Levin Lab
3:20 Stress sharing paper summary
6:15 Paradox of change: Species persist don't evolve
7:20 Bow-tie architectures
10:00  Memories as messages from your past self
12:50 Polycomputing
16:45 Confabulation
17:55 What evidence supports the idea that memories are agential?
22:00 Thought experiment: Entities from earth's core

2. Information Patterns
31:30 Memory is not a filing cabinet
32:30 Are information patterns agential?
35:00  Caterpillar/butterfly... sea slug memory transfer
37:40 Bow-tie architectures are EVERYWHERE
43:20 Bottlenecks "scary" for information

3. Connections & Implications
45:30  Black holes/white holes as bow-ties (Lee Smolin)
47:20 What is confabulation? AI hallucinations
52:30 Gregg Henriques & self-justifying apes... all good agents storytellers
54:20 Information telling stories... Joseph Campbell's journey for a single cell
1:00:50 What comes next?

Works Cited
Self-Improvising Memory: A Perspective on Memories as Agential, Dynamically Reinterpreting Cognitive Glue https://www.mdpi.com/1099-4300/26/6/481
https://thoughtforms.life/suti-the-search-for-unconventional-terrestrial-intelligence/
Biohacking our way to health with robot cells | Michael Levin (Big Think 2023) https://www.youtube.com/watch?v=D35dTp7it_0
https://peregrinecr.com/

What is this channel?
Exploring Truth in philosophy, science, & art. We'll uncover concepts from psychology, mythology, spirituality, literature, media, and more. If you like Lex Fridman or Curt Jaimungal, you'll love this educational channel.

#science #memory #biology #computing #mind #intelligence #attractor #polycomputing #bioelectric #cybernetics #research #life

Voir sur YouTube ↗