Action Schema Networks: Generalised Policies with Deep Learning
This addresses the problem of efficient policy learning in probabilistic planning for AI researchers, offering a domain-general approach with amortised training costs.
The paper tackles the problem of learning generalised policies for probabilistic planning by introducing Action Schema Networks (ASNet), a neural network architecture that mimics relational structures to enable weight-sharing across problems in a domain. The result shows ASNet significantly outperforms traditional non-learning planners in several challenging domains.
In this paper, we introduce the Action Schema Network (ASNet): a neural network architecture for learning generalised policies for probabilistic planning problems. By mimicking the relational structure of planning problems, ASNets are able to adopt a weight-sharing scheme which allows the network to be applied to any problem from a given planning domain. This allows the cost of training the network to be amortised over all problems in that domain. Further, we propose a training method which balances exploration and supervised training on small problems to produce a policy which remains robust when evaluated on larger problems. In experiments, we show that ASNet's learning capability allows it to significantly outperform traditional non-learning planners in several challenging domains.