TY - GEN
T1 - Diversity-based Trajectory and Goal Selection with Hindsight Experience Replay
AU - Dai, Tianhong
AU - Liu, Hengyan
AU - Arulkumaran, Kai
AU - Ren, Guangyu
AU - Bharath, Anil Anthony
PY - 2021
Y1 - 2021
N2 - Hindsight experience replay (HER) is a goal relabelling technique typically used with off-policy deep reinforcement learning algorithms to solve goal-oriented tasks; it is well suited to robotic manipulation tasks that deliver only sparse rewards. In HER, both trajectories and transitions are sampled uniformly for training. However, not all of the agent’s experiences contribute equally to training, and so naive uniform sampling may lead to inefficient learning. In this paper, we propose diversity-based trajectory and goal selection with HER (DTGSH). Firstly, trajectories are sampled according to the diversity of the goal states as modelled by determinantal point processes (DPPs). Secondly, transitions with diverse goal states are selected from the trajectories by using k-DPPs. We evaluate DTGSH on five challenging robotic manipulation tasks in simulated robot environments, where we show that our method can learn more quickly and reach higher performance than other state-of-the-art approaches on all tasks.
AB - Hindsight experience replay (HER) is a goal relabelling technique typically used with off-policy deep reinforcement learning algorithms to solve goal-oriented tasks; it is well suited to robotic manipulation tasks that deliver only sparse rewards. In HER, both trajectories and transitions are sampled uniformly for training. However, not all of the agent’s experiences contribute equally to training, and so naive uniform sampling may lead to inefficient learning. In this paper, we propose diversity-based trajectory and goal selection with HER (DTGSH). Firstly, trajectories are sampled according to the diversity of the goal states as modelled by determinantal point processes (DPPs). Secondly, transitions with diverse goal states are selected from the trajectories by using k-DPPs. We evaluate DTGSH on five challenging robotic manipulation tasks in simulated robot environments, where we show that our method can learn more quickly and reach higher performance than other state-of-the-art approaches on all tasks.
UR - http://dx.doi.org/10.1007/978-3-030-89370-5_3
U2 - 10.1007/978-3-030-89370-5_3
DO - 10.1007/978-3-030-89370-5_3
M3 - Published conference contribution
SN - 9783030893699
SN - 9783030893705
T3 - Lecture Notes in Computer Science
SP - 32
EP - 45
BT - Pacific Rim International Conference on Artificial Intelligence
ER -