Search Results for author: Atsushi Suzuki

Found 6 papers, 0 papers with code

Cumulative Stay-time Representation for Electronic Health Records in Medical Event Time Prediction

no code implementations28 Apr 2022 Takayuki Katsuki, Kohei Miyaguchi, Akira Koseki, Toshiya Iwamori, Ryosuke Yanagiya, Atsushi Suzuki

The MET of non-communicable diseases like diabetes is highly correlated to cumulative health conditions, more specifically, how much time the patient spent with specific health conditions in the past.

Time Series

Generalization Bounds for Graph Embedding Using Negative Sampling: Linear vs Hyperbolic

no code implementations NeurIPS 2021 Atsushi Suzuki, Atsushi Nitanda, Jing Wang, Linchuan Xu, Kenji Yamanishi, Marc Cavazza

Graph embedding, which represents real-world entities in a mathematical space, has enabled numerous applications such as analyzing natural languages, social networks, biochemical networks, and knowledge bases. It has been experimentally shown that graph embedding in hyperbolic space can represent hierarchical tree-like data more effectively than embedding in linear space, owing to hyperbolic space's exponential growth property.

Generalization Bounds Graph Embedding

Generalization Error Bound for Hyperbolic Ordinal Embedding

no code implementations21 May 2021 Atsushi Suzuki, Atsushi Nitanda, Jing Wang, Linchuan Xu, Marc Cavazza, Kenji Yamanishi

Hyperbolic ordinal embedding (HOE) represents entities as points in hyperbolic space so that they agree as well as possible with given constraints in the form of entity i is more similar to entity j than to entity k. It has been experimentally shown that HOE can obtain representations of hierarchical data such as a knowledge base and a citation network effectively, owing to hyperbolic space's exponential growth property.

Riemannian TransE: Multi-relational Graph Embedding in Non-Euclidean Space

no code implementations ICLR 2019 Atsushi Suzuki, Yosuke Enokida, Kenji Yamanishi

Multi-relational graph embedding which aims at achieving effective representations with reduced low-dimensional parameters, has been widely used in knowledge base completion.

Graph Embedding Knowledge Base Completion

Time-Discounting Convolution for Event Sequences with Ambiguous Timestamps

no code implementations6 Dec 2018 Takayuki Katsuki, Takayuki Osogami, Akira Koseki, Masaki Ono, Michiharu Kudo, Masaki Makino, Atsushi Suzuki

This paper proposes a method for modeling event sequences with ambiguous timestamps, a time-discounting convolution.

Time Series

Stable Geodesic Update on Hyperbolic Space and its Application to Poincare Embeddings

no code implementations26 May 2018 Yosuke Enokida, Atsushi Suzuki, Kenji Yamanishi

A hyperbolic space has been shown to be more capable of modeling complex networks than a Euclidean space.

Cannot find the paper you are looking for? You can Submit a new open access paper.