Structured Memory for Neural Turing Machines
This work addresses a specific problem in NTMs for researchers in neural networks, offering incremental improvements to memory organization.
The authors tackled the convergence and overfitting issues in Neural Turing Machines (NTMs) by proposing structured memory organizations, resulting in improved convergence speed and prediction accuracy on copy and associative recall tasks.
Neural Turing Machines (NTM) contain memory component that simulates "working memory" in the brain to store and retrieve information to ease simple algorithms learning. So far, only linearly organized memory is proposed, and during experiments, we observed that the model does not always converge, and overfits easily when handling certain tasks. We think memory component is key to some faulty behaviors of NTM, and better organization of memory component could help fight those problems. In this paper, we propose several different structures of memory for NTM, and we proved in experiments that two of our proposed structured-memory NTMs could lead to better convergence, in term of speed and prediction accuracy on copy task and associative recall task as in (Graves et al. 2014).