Reinforcement Learning Distillation
Online knowledge distillation for multi-environment reinforcement learning.
This work studies how curriculum learning and knowledge distillation can be combined to improve reinforcement learning efficiency across multiple environments. The goal is to transfer useful policy structure during training rather than treating each environment independently.
The original portfolio described this effort as an exploration of efficient reinforcement learning through online distillation. This migration keeps it as a research project entry and leaves room for a future code or paper link.