# S-TRIGGER: Continual State Representation Learning via Self-Triggered Generative Replay

## TL;DR

S-TRIGGER introduces a continual learning method for state representation in control tasks, using self-triggered generative replay to adapt to environment changes without forgetting past knowledge.

## Contribution

It proposes a novel environment change detection and generative replay mechanism for continual state representation learning applicable to VAEs.

## Key findings

- Enables fast, high-performing reinforcement learning with continual state representations.
- Avoids catastrophic forgetting in a continual learning setting.
- Learns autonomously without storing past data, maintaining bounded system size.

## Abstract

We consider the problem of building a state representation model for control, in a continual learning setting. As the environment changes, the aim is to efficiently compress the sensory state's information without losing past knowledge, and then use Reinforcement Learning on the resulting features for efficient policy learning. To this end, we propose S-TRIGGER, a general method for Continual State Representation Learning applicable to Variational Auto-Encoders and its many variants. The method is based on Generative Replay, i.e. the use of generated samples to maintain past knowledge. It comes along with a statistically sound method for environment change detection, which self-triggers the Generative Replay. Our experiments on VAEs show that S-TRIGGER learns state representations that allows fast and high-performing Reinforcement Learning, while avoiding catastrophic forgetting. The resulting system is capable of autonomously learning new information without using past data and with a bounded system size. Code for our experiments is attached in Appendix.