Raphaël Millière: The Vector Grounding Problem and Self-Consciousness

1

Introduction

00:00 • 2min

2

The Origins of AI Research

02:11 • 6min

3

Deep Learning Models and Human Cognitive Capabilities

08:08 • 4min

4

Metaphysics and the Evolution of Human Intelligence

12:11 • 2min

5

The Relationship of Language Models to the World

14:34 • 4min

6

The Importance of Representation in Systems

18:38 • 2min

7

The Relationship Between Conscience and Self-Consciousness

20:29 • 4min

8

The Different Forms of Self-Consciousness

24:11 • 3min

9

The Origins of Self-Representation

26:52 • 4min

10

The Constitutive Importance of Self-Consciousness

30:24 • 2min

11

The Consciousness of Language Models

32:05 • 4min

12

The Impossibility of Consciousness in Language Models

35:49 • 4min

13

The Emergence of Self-Representation in AI

39:52 • 2min

14

The Importance of Graularity in Language Models

41:52 • 4min

15

How to Fine Tune a Language Model to Give Helpful, Harmless, and Honest Answers

45:26 • 3min

16

The Psychological Capacity of Language Models

48:29 • 3min

17

The Vector Grounding Problem Preprint

51:04 • 5min

18

The Grounding Problem in Artificial Language

56:07 • 2min

19

The Grounding Problem With Language Models

58:14 • 2min

20

Different Types of Grounding for Large Language Models

01:00:07 • 4min

21

The Importance of Relational Grounding in Language Learning

01:04:21 • 4min

22

The Distributional Hypothesis of Language Models

01:08:49 • 4min

23

The Influence of Wittgenstein on Language Models

01:12:26 • 3min

24

The Distribution Hypothesis and Language Models

01:15:33 • 3min

25

The Distributional Hypothesis of Large Language Models

01:19:01 • 3min

26

The Possibility of Overcoming the Vector Grounding Problem

01:21:43 • 5min

27

The Importance of Learning From Data in Language Models

01:27:07 • 5min

28

The Role of Teleosimantics in the Exploration of Casual Informational Relations

01:32:03 • 4min

29

The Evolution of Mental State Functions

01:35:49 • 4min

30

The Role of the Telos in Fire Related Behavior

01:40:18 • 3min

31

The Distributional Hypothesis and the Evolution of Language Models

01:43:11 • 3min

32

The Double Dissociation Between Referential Grounding and the Capacity for Multimodality

01:46:03 • 3min

33

The Importance of Multimodality in Language Models

01:49:07 • 3min

34

The Relationship Between Human Cognitive and Deep Learning Models

01:52:12 • 2min

35

The Polarization of Current Discourse and Language Models

01:54:07 • 3min

36

The Semantic Competence of Language Models

01:57:05 • 3min

37

The Emergence of Mechanistic Interpretability

02:00:21 • 5min