Learning Structure from the Ground up---Hierarchical Representation Learning by Chunking

From learning to play the piano to speaking a new language, reusing and recombining previously acquired representations enables us to master complex skills and easily adapt to new environments. Inspired by the Gestalt principle of grouping by proximity and theories of chunking in cognitive science, we propose a hierarchical chunking model (HCM). HCM learns representations from non-i.i.d sequential data from the ground up by first discovering the minimal atomic sequential units as chunks. As learning progresses, a hierarchy of chunk representation is acquired by chunking previously learned representations into more complex representations guided by sequential dependence. We provide learning guarantees on an idealized version of HCM, and demonstrate that HCM learns meaningful and interpretable representations in visual, temporal, visual-temporal domains and language data. Furthermore, the interpretability of the learned chunks enables flexible transfer between environments that share partial representational structure. Taken together, our results show how cognitive science in general and theories of chunking in particular could inform novel and more interpretable approaches to representation learning.