Metatrained Agents Implement Bayesoptimal Agents

# Meta-trained agents implement Bayes-optimal agents

**arXiv ID:** 2010.11223
**Authors:** Vladimir Mikulik, Grégoire Delétang, Tom McGrath, Tim Genewein, Miljan Martic, Shane Legg, Pedro A. Ortega
**Published:** 2020-10-21T18:05:21Z
**Abstract:**
Memory-based meta-learning is a powerful technique to build agents that adapt fast to any task within a target distribution. A previous theoretical study has argued that this remarkable performance is because the meta-training protocol incentivises agents to behave Bayes-optimally. We empirically investigate this claim on a number of prediction and bandit tasks. Inspired by ideas from theoretical computer science, we show that meta-learned and Bayes-optimal agents not only behave alike, but they even share a similar computational structure, in the sense that one agent system can approximately simulate the other. Furthermore, we show that Bayes-optimal agents are fixed points of the meta-learning dynamics. Our results suggest that memory-based meta-learning might serve as a general technique for numerically approximating Bayes-optimal agents - that is, even for task distributions for which we currently don't possess tractable models.

## Skill Description

This skill is generated from the arXiv paper: Meta-trained agents implement Bayes-optimal agents (2010.11223).

## How to Use

[To be filled in by the user or by future automation]

## References

- [arXiv:2010.11223](http://arxiv.org/abs/2010.11223v1)