The Dynamics of Iterated Learning

What is Iterated Learning?

Iterated learning as evolutionary system

Iterated learning models

Main ILM results so far

What can we conclude from ILM results?

Example: structured meaning-spaces

Two problems…

A really simple model

Meanings

Objects as sets of feature-vectors

Big Assumption

Differences…

Simulation

Experiments

Holistic vs. Compositional

Holistic with differing bottlenecks

Differing meaning-spaces

Relative stability for differing meaning spaces (compared with holistic)

Iterated learning with mixed system

Complete dynamics for mixed system

Invention

Effect of low bottleneck

Low bottleneck with invention

Multiple systems

Multiple systems

Multiple systems

Multiple systems

Multiple systems

Multiple systems

Multiple systems

Multiple systems

Multiple systems

Inexpressive meaning-spaces

Problem – no pressure for expressivity

Solution – discrimination and contexts

Selection for contextual discrimination

Summary

What’s next?

Final questions