arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1906.07304cs.LGcs.AIcs.SCstat.ML
收藏

神经引导结构推理

Neurally-Guided Structure Inference

Shanghai Jiao Tong University(上海交通大学);MIT CSAIL(麻省理工学院计算机科学与人工智能实验室);IIIS, Tsinghua University(清华大学智能产业研究院);Department of Brain and Cognitive Sciences, MIT(麻省理工学院脑与认知科学系);Center for Brains, Minds and Machines (CBMM), MIT(麻省理工学院脑、心智与机器中心)

Sidi Lu, Jiayuan Mao, Joshua B. Tenenbaum, Jiajun Wu

摘要:

大多数结构推理方法要么依赖于穷举搜索,要么纯粹是数据驱动的。穷举搜索可以稳健地推断任意复杂数据的结构,但速度较慢。数据驱动的方法允许有效的推理,但当测试数据具有比训练数据更复杂的结构时,不能推广。在本文中,我们提出了一种混合推理算法,神经元引导结构推理(NG-SI),保持基于搜索和数据驱动的方法的优点。NG-SI的关键思想是使用神经网络来指导在结构的组成空间上的分层、逐层搜索。我们评估我们的算法上的两个代表性的结构推理任务:概率矩阵分解和符号程序解析。它在这两项任务上都优于数据驱动和基于搜索的替代方案。

英文摘要:

Most structure inference methods either rely on exhaustive search or are purely data-driven. Exhaustive search robustly infers the structure of arbitrarily complex data, but it is slow. Data-driven methods allow efficient inference, but do not generalize when test data have more complex structures than training data. In this paper, we propose a hybrid inference algorithm, the Neurally-Guided Structure Inference (NG-SI), keeping the advantages of both search-based and data-driven methods. The key idea of NG-SI is to use a neural network to guide the hierarchical, layer-wise search over the compositional space of structures. We evaluate our algorithm on two representative structure inference tasks: probabilistic matrix decomposition and symbolic program parsing. It outperforms data-driven and search-based alternatives on both tasks.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑