Evolutionary Training and Abstraction Yields Algorithmic Generalization of Neural Computers

TL;DR

This paper introduces the Neural Harvard Computer, a memory-augmented neural network architecture that employs abstraction and evolutionary training to learn scalable, transferable, and robust algorithmic solutions with strong generalization across diverse tasks.

Contribution

The paper presents the Neural Harvard Computer, a novel architecture that decouples algorithmic operations from data manipulations, enabling scalable and transferable learning of algorithms.

Findings

Achieves perfect generalization to unseen task configurations.

Learns scalable solutions beyond training complexities.

Independent of data representation and domain.

Abstract

A key feature of intelligent behaviour is the ability to learn abstract strategies that scale and transfer to unfamiliar problems. An abstract strategy solves every sample from a problem class, no matter its representation or complexity -- like algorithms in computer science. Neural networks are powerful models for processing sensory data, discovering hidden patterns, and learning complex functions, but they struggle to learn such iterative, sequential or hierarchical algorithmic strategies. Extending neural networks with external memories has increased their capacities in learning such strategies, but they are still prone to data variations, struggle to learn scalable and transferable solutions, and require massive training data. We present the Neural Harvard Computer (NHC), a memory-augmented network based architecture, that employs abstraction by decoupling algorithmic operations from data manipulations, realized by splitting the information flow and separated modules. This abstraction mechanism and evolutionary training enable the learning of robust and scalable algorithmic solutions. On a diverse set of 11 algorithms with varying complexities, we show that the NHC reliably learns algorithmic solutions with strong generalization and abstraction: perfect generalization and scaling to arbitrary task configurations and complexities far beyond seen during training, and being independent of the data representation and the task domain.