TL;DR本研究提出了一种名为神经随机访问机(Neural Random Access Machine)的神经网络架构,通过反向传播从输入输出示例中训练模型。实验结果表明,该模型能够学习解决需要指针操作和解引用的算法任务,并可以对像链表和二叉树这样的简单数据结构进行操作。
Abstract
In this paper, we propose and investigate a new neural network architecture called neural random access machine. It can manipulate and dereference pointers to an external variable-size random-access memory. The m