Akira Koseki

Title

Senior Research Scientist

Bio

Akira Koseki is a senior researcher at IBM Research - Tokyo. He has been technically leading insurance and healthcare related AI research projects for years.

His research areas include AI techniques to model medical and healthcare proceses such as disease progression from real world electronic health records. His research team has published research papers on disease progression prediction technology for Type 2 Diabetes Mellitus, chronic kidney diseases, and clinical nutrition management.

He has been working for Systems research inclusing middleware performance and Java JIT compiler.

He received a Ph.D. from Waseda University in 1998 for his work on Compiler for Pararrel Architecture. He is a member of IPSJ.

Publications

Improving Health-Status Classification for Epithelial Cells of Inflammatory Bowel Disease using a Transcriptomic Biomedical Foundation Model
- - Toshiya Iwamori
  - Vibha Anand
  - et al.
- 2024
- ISMB 2024
Cross Dataset Verification of Cell Type Annotation using a Transcriptomic Biomedical Foundation Model: Inflammatory Bowel Disease Use Case
- - Akihiro Kosugi
  - Vibha Anand
  - et al.
- 2024
- ISMB 2024
Contextualizing Single-Cell Analyses: An AI Pipeline for Evidence Search from Literature and Gene Databases
- - Joao Bettencourt-Silva
  - Natasha Mulligan
  - et al.
- 2024
- ISMB 2024
Identifying Putative Gene Markers: A Biomedical Foundation Model-based Approach for Cell Type Annotation
- - Uri Kartoun
  - Akira Koseki
  - et al.
- 2024
- ISMB 2024
Effective In-Silico Gene Perturbation by Machine Learning Model Interpretation for Immunotherapies
- - Tanwi Biswas
  - Akira Koseki
  - et al.
- 2024
- ISMB 2024
DPM360: New Additions to Advanced Disease Progression Modeling
- - Akira Koseki
  - Italo Buleje
  - et al.
- 2022
- OHDSI Collaborator Showcase 2022
Variational Inference for Discriminative Learning with Generative Modeling of Feature Incompletion
- - Kohei Miyaguchi
  - Takayuki Katsuki
  - et al.
- 2022
- ICLR 2022
A Comparative Time-to-Event Analysis Across Health Systems
- - Mohamed Ghalwash
  - Prithwish Chakraborty
  - et al.
- 2021
- AMIA Annual Symposium 2021
AI Prediction of Chronic Kidney Disease Stage Progression Combinations for Type 2 Diabetes Mellitus
- - Kohtaroh Miyamoto
  - Akira Koseki
  - et al.
- 2020
- ADA Scientific Sessions 2020
Prediction of heart failures for diabetics only using major longitudinal lab test results
- - Akira Koseki
  - Takuya Ohko
  - et al.
- 2020
- ADA Scientific Sessions 2020

Patents

- 13 Sep 2005
- US
- 6944852
Compiler
- 04 Mar 2005
- JP
- 3651579
A Method To Keep Data-flow Optimizations Effective In The Presence Of Virtual Calls Without Very Large Expansion Of Codesize Or Method Recompilation
- 04 Mar 2005
- JP
- 3651774
A Graph-coloring Method In Which Parallelism Is Taken Into Account Minimally
- 07 Mar 2003
- JP
- 3405696
A Code Generation Technique To Keep Trade-off Between Register And Parallelism Of Program
- 04 Oct 2002
- JP
- 3355602
%inc3%an Optimizing Compilat'n Method Of Processing Multi-dimensarrays For Languages Whose Multi-dimension Arrays Are Implemente

Projects

Java JIT Compiler (at Tokyo)

View more projects

Top collaborators