About us

Consortium

Team

AI Flanders

Contact

Description:
Web application that easily allows a system (robot, machine, computer) to send observations and rewards to a Reinforcement Learning agent, and receive an action in return, all of this over the network using JSON over HTTP requests. Shepherd also comes with an administration page that allows to configure the RL agent, export it for deployment outside Shepherd, and look at its learning curve.

More Information:
<ul>
<li>Grand Challenge/Workpackage(s): GC3 - WP2</li>
<li>Responsible Research Lead: Ann Now&eacute;</li>
<li>Authors: H&eacute;l&egrave;ne Plisnier, Denis Steckelmacher&nbsp;</li>
<li>Programming Language: Python</li>
<li>Link to the repository: <a href="https://github.com/vub-ai-lab/shepherd" target="_blank" rel="noopener">https://github.com/vub-ai-lab/shepherd</a></li>
<li>Link to the website: https://helepli-shepherd-docs.readthedocs.io/</li>
<li>Categories: Reinforcement Learning</li>
</ul>

Reinforcement learning agents as a web service