Learning to Solve Constraint Satisfaction Problems with Recurrent Transformer
This addresses constraint satisfaction problems for AI and reasoning systems, with incremental improvements in handling visual inputs and sample efficiency.
The paper tackled solving constraint satisfaction problems (CSPs) by proposing a Recurrent Transformer that learns end-to-end, showing advantages over state-of-the-art methods like Graph Neural Networks and SATNet, and applied it to visual constraint reasoning to address symbol grounding.
Constraint satisfaction problems (CSPs) are about finding values of variables that satisfy the given constraints. We show that Transformer extended with recurrence is a viable approach to learning to solve CSPs in an end-to-end manner, having clear advantages over state-of-the-art methods such as Graph Neural Networks, SATNet, and some neuro-symbolic models. With the ability of Transformer to handle visual input, the proposed Recurrent Transformer can straightforwardly be applied to visual constraint reasoning problems while successfully addressing the symbol grounding problem. We also show how to leverage deductive knowledge of discrete constraints in the Transformer's inductive learning to achieve sample-efficient learning and semi-supervised learning for CSPs.