RLSS: A Deep Reinforcement Learning Algorithm for Sequential Scene Generation

TL;DR

RLSS introduces a reinforcement learning approach using PPO for sequential scene generation, effectively reducing action space with greedy search, enabling the creation of diverse, plausible scenes aligned with specific design goals.

Contribution

The paper presents a novel RL algorithm for scene generation that combines PPO with greedy search to handle large action spaces and achieve targeted scene properties.

Findings

Converges for a large number of actions

Generates diverse and plausible scenes

Successfully applied to indoor planning and game level design

Abstract

We present RLSS: a reinforcement learning algorithm for sequential scene generation. This is based on employing the proximal policy optimization (PPO) algorithm for generative problems. In particular, we consider how to effectively reduce the action space by including a greedy search algorithm in the learning process. Our experiments demonstrate that our method converges for a relatively large number of actions and learns to generate scenes with predefined design objectives. This approach is placing objects iteratively in the virtual scene. In each step, the network chooses which objects to place and selects positions which result in maximal reward. A high reward is assigned if the last action resulted in desired properties whereas the violation of constraints is penalized. We demonstrate the capability of our method to generate plausible and diverse scenes efficiently by solving indoor planning problems and generating Angry Birds levels.