潜在扩散规划用于模仿学习
Latent Diffusion Planning for Imitation Learning
AI总结：
提出潜在扩散规划（LDP），通过分离规划与动作预测，利用无动作和次优数据，在模拟视觉机器人操作任务上超越现有模仿学习方法。
摘要：
近期模仿学习的进展得益于能够扩展到复杂视觉运动任务、多模态分布和大规模数据集的策略架构。然而，这些方法通常依赖于从大量专家示范中学习。为解决这些不足，我们提出潜在扩散规划（LDP），一种模块化方法，包含一个能够利用无动作示范的规划器和一个能够利用次优数据的逆动力学模型，两者均在学习的潜在空间上运行。首先，我们通过变分自编码器学习一个紧凑的潜在空间，从而在基于图像的领域中有效预测未来状态。然后，我们使用扩散目标训练规划器和逆动力学模型。通过将规划与动作预测分离，LDP能够从次优和无动作数据的更密集监督信号中获益。在模拟视觉机器人操作任务上，LDP优于最先进的模仿学习方法，因为这些方法无法利用此类额外数据。
英文摘要：
Recent progress in imitation learning has been enabled by policy architectures that scale to complex visuomotor tasks, multimodal distributions, and large datasets. However, these methods often rely on learning from large amount of expert demonstrations. To address these shortcomings, we propose Latent Diffusion Planning (LDP), a modular approach consisting of a planner which can leverage action-free demonstrations, and an inverse dynamics model which can leverage suboptimal data, that both operate over a learned latent space. First, we learn a compact latent space through a variational autoencoder, enabling effective forecasting of future states in image-based domains. Then, we train a planner and an inverse dynamics model with diffusion objectives. By separating planning from action prediction, LDP can benefit from the denser supervision signals of suboptimal and action-free data. On simulated visual robotic manipulation tasks, LDP outperforms state-of-the-art imitation learning approaches, as they cannot leverage such additional data.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
