arXiv subjects

Haruo Hosoya

Publications and source records attributed to Haruo Hosoya.

3 recordsLinked to original sources

arXiv2022

CIGMO: Categorical invariant representations in a deep generative framework

Data of general object images have two most common structures: (1) each object of a given shape can be rendered in multiple different views, and (2) shapes of objects can be categorized in such a way that the diversity of shapes is much larger across categories than within a category. Existing deep generative models can typically capture either structure, but not both. In this work, we introduce a novel deep generative model, called CIGMO, that can learn to represent category, shape, and view factors from image data. The model is comprised of multiple modules of shape representations that are each specialized to a particular category and disentangled from view representation, and can be learned using a group-based weakly supervised learning method. By empirical investigation, we show that our model can effectively discover categories of object shapes despite large view variation and quantitatively supersede various previous methods including the state-of-the-art invariant clustering algorithm. Further, we show that our approach using category-specialization can enhance the learned shape representation to better perform down-stream tasks such as one-shot object identification as well as shape-view disentanglement.

cs.CV↗

arXiv2021

Group-based Learning of Disentangled Representations with Generalizability for Novel Contents

Sensory data are often comprised of independent content and transformation factors. For example, face images may have shapes as content and poses as transformation. To infer separately these factors from given data, various ``disentangling'' models have been proposed. However, many of these are supervised or semi-supervised, either requiring attribute labels that are often unavailable or disallowing for generalization over new contents. In this study, we introduce a novel deep generative model, called group-based variational autoencoders. In this, we assume no explicit labels, but a weaker form of structure that groups together data instances having the same content but transformed differently; we thereby separately estimate a group-common factor as content and an instance-specific factor as transformation. This approach allows for learning to represent a general continuous space of contents, which can accommodate unseen contents. Despite the simplicity, our model succeeded in learning, from five datasets, content representations that are highly separate from the transformation representation and generalizable to data with novel contents. We further provide detailed analysis of the latent content code and show insight into how our model obtains the notable transformation invariance and content generalizability.

cs.LG↗

arXiv2007

Towards Practical Typechecking for Macro Tree Transducers

Macro tree transducers (mtt) are an important model that both covers many useful XML transformations and allows decidable exact typechecking. This paper reports our first step toward an implementation of mtt typechecker that has a practical efficiency. Our approach is to represent an input type obtained from a backward inference as an alternating tree automaton, in a style similar to Tozawa's XSLT0 typechecking. In this approach, typechecking reduces to checking emptiness of an alternating tree automaton. We propose several optimizations (Cartesian factorization, state partitioning) on the backward inference process in order to produce much smaller alternating tree automata than the naive algorithm, and we present our efficient algorithm for checking emptiness of alternating tree automata, where we exploit the explicit representation of alternation for local optimizations. Our preliminary experiments confirm that our algorithm has a practical performance that can typecheck simple transformations with respect to the full XHTML in a reasonable time.

cs.PL↗