Joel Lehman: Open-Endedness and Evolution through Large Models

1

Introduction

00:00 • 3min

2

Evolutionary Algorithms - What Drew You to Them?

03:09 • 2min

3

Is This a Volitionless Process?

04:52 • 4min

4

The Search for Interestingness in Evolutionary Algorithms

08:40 • 4min

5

The Concept of Deception in Machine Learning

12:52 • 2min

6

Is There a Reward Shift in a Novelty Search?

14:44 • 3min

7

The Intuition of Novel Research Circumvents Deception

17:20 • 3min

8

A Sense of Humility in Deep Learning and Neural Networks

19:54 • 3min

9

Using a Measure Post Hoc to Optimize for the Thing You Really Care About

22:46 • 2min

10

Is There a Desire to Prevent Premature Convergence to Local Optimization?

24:49 • 2min

11

Neural Evolution Algorithms and the Evolution of Virtual Creatures

26:26 • 3min

12

The Novel Research Oven - What Is It About?

29:02 • 3min

13

The Combination of Fitness and Novelty

32:31 • 4min

14

The Relationship Between Evolution and Deep Learning

36:02 • 5min

15

Neural Evolution in Deep Learning

41:13 • 4min

16

Neural Networks Are a Disembodied Paradigm

45:40 • 2min

17

The Color Red Is a Very Interesting Epistemological Question

48:05 • 2min

18

Is There a Synergy Between Language Models and Evolution?

50:14 • 5min

19

Is This a Challenge in Open-Endedness?

55:22 • 5min

20

ELM in a Quality Diversity Algorithm

01:00:07 • 3min

21

Is There a Difference Between the Scale of the Grid and the Dimensions of Diversity of the Map Elite Script?

01:03:29 • 4min

22

I Think Interestingness Is Something Subjective or There's an Objective Component to It

01:07:42 • 2min

23

The Three-Stage Invention Pipeline

01:10:11 • 4min

24

Is Scaling a Good Idea?

01:14:07 • 2min

25

What Did You Find Most Interesting About the Experiments You Did?

01:16:35 • 2min

26

What Are Some of the Key Challenges in Open Edness?

01:18:37 • 4min

27

The Myth of the Objective

01:22:41 • 4min

28

Is There a Difference Between a Novelty Search and an Uber AI?

01:26:18 • 2min

29

Is There a Happy Medium?

01:28:35 • 4min

30

Is There a Mentorship in Machine Learning?

01:32:35 • 5min

31

The Gradient on Sub Stack Podcast

01:37:15 • 2min