Hangul Fonts Dataset: a Hierarchical and Compositional Dataset for Investigating Learned Representations

• Main Authors: Livezey, Jesse A, Hwang, Ahyeon, Yeung, Jacob, Bouchard, Kristofer E
• Format: Journal Article
• Language: English
• Published: 23-05-2019
• Subjects: Computer Science - Computer Vision and Pattern Recognition; Computer Science - Learning; Computer Science - Neural and Evolutionary Computing
• DOI: 10.48550/arxiv.1905.13308

Hierarchy and compositionality are common latent properties in many natural
and scientific datasets. Determining when a deep network's hidden activations
represent hierarchy and compositionality is important both for understanding
deep representation learning and for applying deep networks in domains where
interpretability is crucial. However, current benchmark machine learning
datasets either have little hierarchical or compositional structure, or the
structure is not known. This gap impedes precise analysis of a network's
representations and thus hinders development of new methods that can learn such
properties. To address this gap, we developed a new benchmark dataset with
known hierarchical and compositional structure. The Hangul Fonts Dataset (HFD)
is comprised of 35 fonts from the Korean writing system (Hangul), each with
11,172 blocks (syllables) composed from the product of initial consonant,
medial vowel, and final consonant glyphs. All blocks can be grouped into a few
geometric types which induces a hierarchy across blocks. In addition, each
block is composed of individual glyphs with rotations, translations, scalings,
and naturalistic style variation across fonts. We find that both shallow and
deep unsupervised methods only show modest evidence of hierarchy and
compositionality in their representations of the HFD compared to supervised
deep networks. Supervised deep network representations contain structure
related to the geometrical hierarchy of the characters, but the compositional
structure of the data is not evident. Thus, HFD enables the identification of
shortcomings in existing methods, a critical first step toward developing new
machine learning algorithms to extract hierarchical and compositional structure
in the context of naturalistic variability.