> ML_LITERATURE // LILLICRAP-2016-CONTINUOUS-CONTROL-WITH-DEEP-REINFORCEMENT-LEARNING-DDPG_v1.0
Continuous control with deep reinforcement learning (DDPG)
Timothy P. Lillicrap, Jonathan J. Hunt, Alexander Pritzel, Nicolas Heess, Tom Erez, Yuval Tassa, David Silver, Daan Wierstra · International Conference on Learning Representations (ICLR) (2016)
algorithm2016foundationalthirdPartyReproduced
Principal Contribution
Adapted Deterministic Policy Gradient (DPG) to continuous action spaces using deep neural networks with actor-critic architecture, replay memory, and soft target updates.
Operational Relevance
Serves as qualified theoretical and systems foundation for task-reinforcement-learning, task-continuous-control.
Assumptions
- Markovian state dynamics and stationary reward functions hold in target evaluation environments
Limitations
- Sample efficiency, exploration stability, and real-world sim-to-real transfer gaps require specialized tuning
Connected Algorithms, Architectures & Tools
Related Algorithms:
Related Architectures:
Implementing Libraries: