Platform adds variational option discovery algorithms

AI Impact Summary

This release exposes variational option discovery capabilities, enabling agents to identify useful higher-level actions via variational inference. The addition can improve sample efficiency and planning horizon handling for RL and sequential decision tasks, reducing training time and enabling more scalable policies. Technical teams should assess current RL pipelines for compatibility with variational wrappers, ensure resource provisioning for probabilistic models and optimization, and plan validation experiments to benchmark the new algorithms.