# An Optimization Framework for Task Sequencing in Curriculum Learning

## TL;DR

This paper introduces a general optimization framework for task sequencing in curriculum learning within reinforcement learning, demonstrating its ability to enhance initial performance, reduce exploration costs, and find superior policies.

## Contribution

It proposes a novel, flexible framework for task sequencing based on different objectives and evaluates metaheuristic search methods for optimizing curriculum sequences.

## Key findings

- Improved initial performance in reinforcement learning agents
- Fewer suboptimal actions during exploration
- Discovery of better policies through curriculum sequencing

## Abstract

Curriculum learning in reinforcement learning is used to shape exploration by presenting the agent with increasingly complex tasks. The idea of curriculum learning has been largely applied in both animal training and pedagogy. In reinforcement learning, all previous task sequencing methods have shaped exploration with the objective of reducing the time to reach a given performance level. We propose novel uses of curriculum learning, which arise from choosing different objective functions. Furthermore, we define a general optimization framework for task sequencing and evaluate the performance of popular metaheuristic search methods on several tasks. We show that curriculum learning can be successfully used to: improve the initial performance, take fewer suboptimal actions during exploration, and discover better policies.