Abstract Reasoning via Logic-guided Generation
This addresses the challenge of improving generative abstract reasoning for AI systems, representing an incremental advance over prior discriminative approaches.
The paper tackles abstract reasoning by proposing a generative deep neural network framework called logic-guided generation (LoGe), which reduces the problem to propositional logic optimization and outperforms black-box methods on the RAVEN benchmark.
Abstract reasoning, i.e., inferring complicated patterns from given observations, is a central building block of artificial general intelligence. While humans find the answer by either eliminating wrong candidates or first constructing the answer, prior deep neural network (DNN)-based methods focus on the former discriminative approach. This paper aims to design a framework for the latter approach and bridge the gap between artificial and human intelligence. To this end, we propose logic-guided generation (LoGe), a novel generative DNN framework that reduces abstract reasoning as an optimization problem in propositional logic. LoGe is composed of three steps: extract propositional variables from images, reason the answer variables with a logic layer, and reconstruct the answer image from the variables. We demonstrate that LoGe outperforms the black box DNN frameworks for generative abstract reasoning under the RAVEN benchmark, i.e., reconstructing answers based on capturing correct rules of various attributes from observations.