因果动作影响感知的反事实数据增强
Causal Action Influence Aware Counterfactual Data Augmentation
AI总结：
针对离线机器人学习中数据稀缺易致策略学习非因果关系的问题，提出CAIAC反事实数据增强方法，通过量化因果影响交换轨迹间动作无关状态部分生成合成数据，提升离线算法抗分布偏移鲁棒性。
摘要：
离线数据是教会机器人复杂行为的宝贵且实用的资源。理想情况下，学习智能体不应受限于可用演示数据的稀缺性，而应能在训练分布之外实现泛化。然而，现实场景的复杂性通常需要海量数据，以防止神经网络策略捕捉到虚假相关性并学习到非因果关系。我们提出CAIAC（因果动作影响感知反事实数据增强），这是一种无需在线环境交互即可从固定数据集生成可行合成转换的数据增强方法。通过利用量化因果影响的原则性方法，我们能够通过在数据集的独立轨迹之间交换状态空间中不受动作影响的部分来开展反事实推理。我们通过实验证明，该方法可大幅提升离线学习算法应对分布偏移的鲁棒性。
英文摘要：
Offline data are both valuable and practical resources for teaching robots complex behaviors. Ideally, learning agents should not be constrained by the scarcity of available demonstrations, but rather generalize beyond the training distribution. However, the complexity of real-world scenarios typically requires huge amounts of data to prevent neural network policies from picking up on spurious correlations and learning non-causal relationships. We propose CAIAC, a data augmentation method that can create feasible synthetic transitions from a fixed dataset without having access to online environment interactions. By utilizing principled methods for quantifying causal influence, we are able to perform counterfactual reasoning by swapping $\it{action}$-unaffected parts of the state-space between independent trajectories in the dataset. We empirically show that this leads to a substantial increase in robustness of offline learning algorithms against distributional shift.
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。
