Semantic Scholar Open Access 2015 15157 sitasi

Continuous control with deep reinforcement learning

T. Lillicrap Jonathan J. Hunt A. Pritzel N. Heess Tom Erez +3 lainnya

Abstrak

We adapt the ideas underlying the success of Deep Q-Learning to the continuous action domain. We present an actor-critic, model-free algorithm based on the deterministic policy gradient that can operate over continuous action spaces. Using the same learning algorithm, network architecture and hyper-parameters, our algorithm robustly solves more than 20 simulated physics tasks, including classic problems such as cartpole swing-up, dexterous manipulation, legged locomotion and car driving. Our algorithm is able to find policies whose performance is competitive with those found by a planning algorithm with full access to the dynamics of the domain and its derivatives. We further demonstrate that for many of the tasks the algorithm can learn policies end-to-end: directly from raw pixel inputs.

Penulis (8)

T

T. Lillicrap

J

Jonathan J. Hunt

A

A. Pritzel

N

N. Heess

T

Tom Erez

Y

Yuval Tassa

D

David Silver

D

Daan Wierstra

Format Sitasi

Lillicrap, T., Hunt, J.J., Pritzel, A., Heess, N., Erez, T., Tassa, Y. et al. (2015). Continuous control with deep reinforcement learning. https://www.semanticscholar.org/paper/024006d4c2a89f7acacc6e4438d156525b60a98f

Akses Cepat

PDF tidak tersedia langsung

Cek di sumber asli →
Lihat di Sumber
Informasi Jurnal
Tahun Terbit
2015
Bahasa
en
Total Sitasi
15157×
Sumber Database
Semantic Scholar
Akses
Open Access ✓