How to represent part-whole hierarchies in a neural network
This addresses the challenge of interpretable hierarchical parsing in AI, but it is incremental as it combines existing advances without a working system.
The paper tackles the problem of representing part-whole hierarchies in neural networks with fixed architectures, proposing an idea called GLOM that uses islands of identical vectors to represent parse tree nodes, which could improve interpretability in vision or language systems.
This paper does not describe a working system. Instead, it presents a single idea about representation which allows advances made by several different groups to be combined into an imaginary system called GLOM. The advances include transformers, neural fields, contrastive representation learning, distillation and capsules. GLOM answers the question: How can a neural network with a fixed architecture parse an image into a part-whole hierarchy which has a different structure for each image? The idea is simply to use islands of identical vectors to represent the nodes in the parse tree. If GLOM can be made to work, it should significantly improve the interpretability of the representations produced by transformer-like systems when applied to vision or language