INTRODUCTION

Creating a world where everyone can do what they want through the MUSUBI project.

To achieve "human understanding and respect" and "diversity," artificial intelligence technologies that can analyze not only individual attributes but also human relationships are necessary. Graph analysis is widely used to analyze relationships in applications such as social network analysis in SNS users, product recommendations in web shopping platforms like Amazon, and the structuring of knowledge in Google search.

The goal of this research project is to develop MUSUBI framework for graph analysis. We believe that advancing and democratizing graph analysis is necessary to understand our world and to create a world where everyone can do what they want.

1. Advancing Graph Analysis

Graph data is often used to model relationships as data. Graph analysis techniques need to be developed from multiple perspectives. Key challenges include prediction of labels and links within a graph, ranking node importance, detection of patterns, and data partitioning. To address these challenges, we aim to develop deep learning techniques, node importance analysis methods, clustering techniques, and patten detection methods for graph data. These advancements will enable more effective graph data analysis.

We do not focus on the utility of graph analysis but also fairness. Fair graph data analysis is expected to develop in various fields, including sociological analysis of human relationships or the relationships between people and professions, as well as industrial applications such as product recommendations in e-commerce. We deepen fair graph analysis.

view more

2. Benchmarking and Tool Development for Graph Data Analysis

To evaluate the performance of graph data analysis, we will construct real-world graphs such as knowledge graphs, SNS data, citation networks, and academic collaboration data. Additionally, we will benchmark existing technologies through automation and graph generation techniques. By implementing a framework that facilitates graph deep learning and data mining, we aim to democratize both graph analysis and fair graph analysis.

view more

3. Multidimensional Analysis of Fairness in Real-World Graphs and AI Technologies

We will evaluate and clarify the social biases present in real-world graph data from a data science perspective. Furthermore, we will assess the fairness of AI technologies through the lens of philosophy and ethics, examining their societal impact and philosophical implications. Additionally, we will explore the necessity of new evaluation metrics for fairness in graph analysis. Through this approach, we aim to pursue fairness in real-world graphs and AI technologies from both data science and philosophical-ethical perspectives.

view more

MEMBER

PI

Yuya Sasaki

Computer Science

Osaka University

Co-PI

Kazuki Nakajima

Network Science

Tokyo Metropolitan Unversity

Co-PI

George Fletcher

Computer Science

Techinical University of Eindhoven

Co-PI

Panagiotis Karras

Computer Science

The University of Copenhagen

Onizuka Makoto

Computer Science

Osaka University

Daichi Amagata

Computer Science

Osaka University

Sohei Tokuno

Philosophy

NAIST

Osamu Sakura

Science and technology studies

The University of Tokyo

Haruka Maeda

Social Science

Kyoto University

Nicklay Yakovets

Computer Science

Techinical University of Eindhoven

Mykola Pechenizkiy

Computer Science

Techinical University of Eindhoven

Irina Shklovski

Computer Science

The University of Copenhagen

Natalia Avlona

Law & Computer Science

The University of Copenhagen

Eiichiro Watamura

Psycology

Osaka University

Boris Düdder

Computer Science

The University of Copenhagen

Xikun Jiang

Computer Science

The University of Copenhagen

student

Taiki Ishisaki

Computer Science

Osaka University

student

Konstantinos Skitsas

Computer Science

The University of Copenhagen

student

He Lyu

Computer Science

Osaka University

student

Yutaka Kojima

Computer Science

Osaka University

NEWS

March 27, 2025 Paper

Paper Accepted at ICDE 2025!

Daich Amagata will present his work related to fairness-aware join at ICDE 2025 held in Hong Kong! See you there!

Daichi Amagata, "Random Sampling over Spatial Range Joins", In Proceedings of the IEEE International Conference on Data Engineering (ICDE), 2025.

view more

March 26, 2025 Event

ASPIRE annual meeting

There is an ASPIRE annual meeting with other PIs, organizers, and advisors in Tokyo.
Yuya reported our progress and discussed further research collaborations!

view more

March 25, 2025 Event

Talk at Complex Systems Research Exchange

Yuya was invited to talk about our Musubi project, "Fairness in graph data analysis," at Complex Systems Research Exchange (CREx).
Complex Systems Research Exchange (CREx) is an online seminar series aimed at building an international research community in the interdisciplinary field of complex systems.
https://sites.google.com/view/cxrex/home

We are happy to talk with people in various fields!

view more

March 03, 2025 Event

Tutorial talk at DEIM 2025

Yuya did a tutorial talk, "An Encouragement of Fairness-aware Graph Analysis," at DEIM 2025.
DEIM is the biggest database conference in Japan.

view more

February 28, 2025 Tool

MUSUBI website open!

We open our website!

view more

February 01, 2025 Tool

The Beginning of Evie' Long-term Stay at Osaka University

Evie who is a master student at Tue, Netherland, starts to study in Osaka university. She aims to build benchmarks for query visualizations!

view more

January 30, 2025 Meeting

Yuya visited Tue

Yuya visited Tue from 10th to 20th January 2025.
George, Mykola, Yuya, and new students discussed new research directions!

view more

December 16, 2024 Event

Tutorial talk at ADC 2024

Yuya did a tutorial talk, "Fairness in graph data analysis", at the Australian Database Conference (ADC) 2025.

view more

December 10, 2024 Tool

Kyoto University Tokyo Office meeting

Daichi (online), Haruka, Makoto, Osamu, Yutaka, and Yuya discussed the 2024 year' progress of our research, at Kyoto University Tokyo Office!

view more

November 01, 2024 Event

6th KJMD workshop

We held and joined KJMD which is a joint workshop with Postec, south Korea, and Osaka university, in Hokaiddo, Japan!
Our students talked about their studies for database, machine learning, and graph analysis.
https://sites.google.com/view/kjmd2024/

view more

September 15, 2024 Event

Shonan meeting

George, Mykola, Sohei, and Yuya joined Shonan meeting "Understanding the “Why” of Data and Knowledge Models" from 15th to 20th September.

view more

September 07, 2024 Meeting

Yuya visits Copenhagen

Yuya visited the University of Copenhagen and discussed new research directions with Irina, Natalia, Boris, and Xikun!

view more

August 29, 2024 Tool

Logo and goods

We made our logo and goods for MUSUBI project, which are very cool Japanese-traditional style!
Please join our MUSUBI project to use and get them!

view more

July 28, 2024 Meeting

Visiting IJCAI 2024 and Osaka university

Mykola, Panos, Konstantinos, and Yuya visited IJCAI 2024 together, held on Jeju island, south Korea.
Mykola did a tutorial about fair reinforcement learning, which was super interesting! https://fair-rl.github.io/
Before IJCAI, Mykola (and his family) and Panos stayed at Osaka University. We really enjoyed discussions!

view more

July 27, 2024 Meeting

The Beginning of Konstantinos' Long-term Stay at Osaka University

Konstantinos, who is a PhD student at Aarhus University, Denmark, stays at Osaka University until the end of December.

view more

July 24, 2024 Event

2nd Osaka Trustworthy AI workshop

We held the 2nd Osaka Trustworthy AI workshop, and Mykola gave a talk about "On Measuring and Mitigating Unfairness in Algorithmic Decision Making".

view more

May 22, 2024 Event

Tue Database group lunch talk

Yuya talked about our MUSUBI project in Tue on 22nd May.

view more

May 13, 2024 Meeting

ICDE 2024 and Visiting Tue

Daich, George, and Yuya participated in ICDE 2024 held in Utrecht, Netherlands.
Daich made a presentation about his paper "Independent Range Sampling on Interval Data".
After ICDE, Yuya visited Tue from 13th to 29th May 2024. We discussed our further collaborations.

view more

March 11, 2023 Event

1st Osaka Trustworthy AI workshop

We held the 1st Osaka Trustworthy AI workshop. We invited 7 researchers who actively study the trustworthy of AIs.
From MUSUBI project, Daich and Yuya gave talks about "Evaluating Fairness Metrics Across Borders from Human Perceptions" and "Fair Clustering & Search", respectively.

view more

February 28, 2023 Meeting

kick-off meeting

We held our kick-off meeting on hybrid style. Our project starts actively from here!

view more

PUBLICATION

Book

AI and Courts in Japan

Eiichiro Watamura.

The Cambridge Handbook of AI and Technologies in Courts, chapter 31

Conference

Random Sampling over Spatial Range Joins

Daichi Amagata.

Proceedings of the IEEE International Conference on Data Engineering (ICDE), 2025.

Conference

NAAM: Node-Aware Attention Mechanism for Distilling GNNs-to-MLP

Itsuki Nakayama, Makoto Onizuka.

The 39th Annual AAAI Conference on Artificial Intelligence (AAAI) student abstract, 2025.

Journal

Efficient intervention in the spread of misinformation in social networks

Takumi Sakiyama, Kazuki Nakajima, Masaki Aida.

IEEE Access. Vol. 12, pp. 133489-133498, 2024.

Arrow
Arrow

Conference

Quantifying Gendered Citation Imbalance in Computer Science Conference

Kazuki Nakajima, Yuya Sasaki, Sohei Tokuno, George Fletcher.

Proceedings of 7th AAAI Conference on AI, Ethics, and Society (AIES), 2024.

Arrow

Conference

Mining Path Association Rules in Large Property Graphs

Yuya Sasaki, Panagiotis Karras.

Proceedings of 33rd ACM International Conference on Information and Knowledge Management (CIKM), 2024.

Arrow

Conference

Independent Range Sampling on Interval Data

Daichi Amagata.

Proceedings of The IEEE 40th International Conference on Data Engineering (ICDE), pp. 449-461, 2024

Arrow

Conference

Fair k-center Clustering with Outliers

Daichi Amagata.

Proceedings of The 27th International Conference on Artificial Intelligence and Statistics (AISTAT), pp. 10-18, 2024.

Arrow

Conference

Inference and visualization of community structure in grant collaboration hypergraphs

Kazuki Nakajima, Takeaki Uno.

The 13th International Conference on Complex Networks and Their Applications, 2024.

Conference

Estimating hyperedge size distribution via random walk on hypergraphs

Masanao Kodakari, Kazuki Nakajima, and Masaki Aida.

The 13th International Conference on Complex Networks and Their Applications, 2024.