Implicit planning has emerged as an elegant technique for combining learned models of the world with end-to-end model-free reinforcement learning. We study the class of implicit planners inspired by value iteration, an algorithm that is guaranteed to yield perfect policies in fully-specified tabular environments. We find that prior approaches either assume that the environment is provided in such a tabular form -- which is highly restrictive -- or infer "local neighbourhoods" of states to run value iteration over -- for which we discover an algorithmic bottleneck effect. This effect is caused by explicitly running the planning algorithm based on scalar predictions in every state, which can be harmful to data efficiency if such scalars are improperly predicted. We propose eXecuted Latent Value Iteration Networks (XLVINs), which alleviate the above limitations. Our method performs all planning computations in a high-dimensional latent space, breaking the algorithmic bottleneck. It maintains alignment with value iteration by carefully leveraging neural graph-algorithmic reasoning and contrastive self-supervised learning. Across eight low-data settings -- including classical control, navigation and Atari -- XLVINs provide significant improvements to data efficiency against value iteration-based implicit planners, as well as relevant model-free baselines. Lastly, we empirically verify that XLVINs can closely align with value iteration.
翻译:隐性规划已经成为一种优雅的技术,将世界已学的模型与端到端的无模型强化学习相结合。我们研究了受价值迭代启发的隐含规划者阶级,这种算法保证在完全指定的表格环境中产生完美的政策。我们发现,先前的方法要么假定环境是以列表形式提供的,这种表格形式非常严格,要么推断为“本地街区”进行增值迭代,为此我们发现一种算法瓶颈效应。这种效果是由于在每一个州根据星标预测进行明确的规划算法造成的,如果不适当地预测,这种算法可能对数据效率有害。我们建议采用经裁剪切的点值迭代网络(XLINs),这种算法可以减轻上述限制。我们的方法是在高空潜藏空间进行所有规划计算,打破算法瓶颈。我们通过仔细利用神经图形-横向推理和对比自我监控学习,在八个低数据环境环境环境中,包括古典控制、导航和AtLSLISIN系统,以及我们作为相关基线的精确度数据,提供与无损性数据对比的升级。