Learning Domain-Independent Heuristics for Grounded and Lifted Planning
This work addresses the challenge of scalable and efficient planning in AI, offering a domain-independent solution that improves over prior methods, though it is incremental in advancing graph-based heuristic learning.
The authors tackled the problem of learning domain-independent heuristics for planning tasks by introducing three novel graph representations using Graph Neural Networks, with a method that uses only lifted representations to avoid issues from large grounded models. Their experiments showed that these heuristics generalize to much larger problems than those in training, vastly surpassing the existing STRIPS-HGN heuristics.
We present three novel graph representations of planning tasks suitable for learning domain-independent heuristics using Graph Neural Networks (GNNs) to guide search. In particular, to mitigate the issues caused by large grounded GNNs we present the first method for learning domain-independent heuristics with only the lifted representation of a planning task. We also provide a theoretical analysis of the expressiveness of our models, showing that some are more powerful than STRIPS-HGN, the only other existing model for learning domain-independent heuristics. Our experiments show that our heuristics generalise to much larger problems than those in the training set, vastly surpassing STRIPS-HGN heuristics.