论文标题
学会用回忆的唤醒学习来学习生成程序
Learning to learn generative programs with Memoised Wake-Sleep
论文作者
论文摘要
我们研究了一类神经符号生成模型，其中神经网络既用于推论，又用作符号生成的程序的先验。作为生成模型，这些程序以自然解释的形式捕获组成结构。为了应对学习计划的挑战，作为学习的“内在环”，我们提出了记忆的唤醒（MWS）算法，该算法通过在整个培训中明确存储和重复推理网络发现的最佳程序来扩展唤醒睡眠。我们使用MW在三个具有挑战性的领域中学习准确的，可解释的模型：基于中风的角色建模，蜂窝自动机和几乎没有射击的学习在真实世界字符串概念的新数据集中。
We study a class of neuro-symbolic generative models in which neural networks are used both for inference and as priors over symbolic, data-generating programs. As generative models, these programs capture compositional structures in a naturally explainable form. To tackle the challenge of performing program induction as an 'inner-loop' to learning, we propose the Memoised Wake-Sleep (MWS) algorithm, which extends Wake Sleep by explicitly storing and reusing the best programs discovered by the inference network throughout training. We use MWS to learn accurate, explainable models in three challenging domains: stroke-based character modelling, cellular automata, and few-shot learning in a novel dataset of real-world string concepts.
