Cogment: Open Source Framework For Distributed Multi-actor Training, Deployment & Operations

TL;DR

Cogment is an open-source framework that facilitates scalable, distributed multi-actor training involving humans and AI agents, addressing architectural and operational complexities in human-in-the-loop reinforcement learning.

Contribution

It introduces an actor formalism and a microservice architecture to unify and simplify human-AI collaboration and training workflows.

Findings

Supports various human-in-the-loop training paradigms

Enables scalable distributed training out of the box

Addresses architectural and operational complexities

Abstract

Involving humans directly for the benefit of AI agents' training is getting traction thanks to several advances in reinforcement learning and human-in-the-loop learning. Humans can provide rewards to the agent, demonstrate tasks, design a curriculum, or act in the environment, but these benefits also come with architectural, functional design and engineering complexities. We present Cogment, a unifying open-source framework that introduces an actor formalism to support a variety of humans-agents collaboration typologies and training approaches. It is also scalable out of the box thanks to a distributed micro service architecture, and offers solutions to the aforementioned complexities.