Obstacle Tower
Introduced by Juliani et al. in Obstacle Tower: A Generalization Challenge in Vision, Control, and Planning
Obstacle Tower is a high fidelity, 3D, 3rd person, procedurally generated environment for reinforcement learning. An agent playing Obstacle Tower must learn to solve both low-level control and high-level planning problems in tandem while learning from pixels and a sparse reward signal. Unlike other benchmarks such as the Arcade Learning Environment, evaluation of agent performance in Obstacle Tower is based on an agent’s ability to perform well on unseen instances of the environment.
Source: Obstacle Tower: A Generalization Challenge in Vision, Control, and PlanningBenchmarks
| Trend | Task | Dataset Variant | Best Model | Paper | Code |
|---|---|---|---|---|---|
|
Obstacle Tower (No Gen) fixed
|
RNB
|
|||
|
Obstacle Tower (No Gen) varied
|
RNB
|
|||
|
Obstacle Tower (Weak Gen) fixed
|
PPO
|
|||
|
Obstacle Tower (Weak Gen) varied
|
RNB
|
|||
|
Obstacle Tower (Strong Gen) fixed
|
PPO
|
|||
|
Obstacle Tower (Strong Gen) varied
|
RNB
|
Papers
| Paper | Code | Results | Date | Stars |
|---|
