Multigrid Neural Memory

TL;DR

This paper presents a multigrid neural network architecture with internal, distributed memory that enables long-term, large-scale memory retention and efficient data routing, outperforming prior external memory approaches.

Contribution

The authors introduce a novel multigrid neural network with hierarchical, distributed internal memory, simplifying memory mechanisms while enabling emergent, large-scale, long-term memory capabilities.

Findings

Effective long-term memory retention over thousands of steps.

Superior performance on exploration, mapping, sorting, and question answering tasks.

Memory self-organization and dynamic data-dependent read/write capabilities.

Abstract

We introduce a novel approach to endowing neural networks with emergent, long-term, large-scale memory. Distinct from strategies that connect neural networks to external memory banks via intricately crafted controllers and hand-designed attentional mechanisms, our memory is internal, distributed, co-located alongside computation, and implicitly addressed, while being drastically simpler than prior efforts. Architecting networks with multigrid structure and connectivity, while distributing memory cells alongside computation throughout this topology, we observe the emergence of coherent memory subsystems. Our hierarchical spatial organization, parameterized convolutionally, permits efficient instantiation of large-capacity memories, while multigrid topology provides short internal routing pathways, allowing convolutional networks to efficiently approximate the behavior of fully connected networks. Such networks have an implicit capacity for internal attention; augmented with memory, they learn to read and write specific memory locations in a dynamic data-dependent manner. We demonstrate these capabilities on exploration and mapping tasks, where our network is able to self-organize and retain long-term memory for trajectories of thousands of time steps. On tasks decoupled from any notion of spatial geometry: sorting, associative recall, and question answering, our design functions as a truly generic memory and yields excellent results.