Keynote 1
Session Chair: Wook-Shin Han (POSTECH)
What Can Cognitive Psychology Do For Data Management?
Keynote 1
Session Chair: Wook-Shin Han (POSTECH)
What Can Cognitive Psychology Do For Data Management?
Abstract
Imagine a scenario where we could communicate with a relational query optimizer through messaging platforms like WeChat or WhatsApp, asking questions related to query processing and optimization. If realized, such a tool could prove invaluable in database administration and education, among others. In this talk, we share our ongoing efforts to develop a user-friendly tool, ChatQPT, which facilitates chat- based interactions with a relational query engine (PostgreSQL). We will specifically highlight the complex challenges encountered in addressing this problem and our initial approaches to overcoming them.
Bio
Sourav S. Bhowmick is an Associate Professor in the College of Computing & Data Science (CCDS), Nanyang Technological University, Singapore. His research interests are in data management, human-data interaction, and data analytics. His research has appeared in premium venues such as ACM SIGMOD, VLDB, and VLDB Journal. He is co-recipient of Best Paper Awards in ACM CIKM 2004, ACM BCB 2011, VLDB 2021, and ER 2023. He is also co-recipient of the ACM SIGMOD Research Highlights Award in 2022. Sourav is serving as a member of the SIGMOD Executive Committee, SIGMOD Awards Committee, a regular member of the PVLDB advisory board, and an elected trustee of the VLDB Endowment. He is a co-recipient of several service awards including VLDB Service Award in 2018, Distinguished AE Award in SIGMOD 2021, SIGMOD 2023 and VLDB 2022, and Distinguished Reviewer Award in VLDB 2020, VLDB 2023, and VLDB 2025. He was inducted into Distinguished Members of the ACM in 2020. Sourav is a strong advocate of research that directly or indirectly impacts end users.
Keynote 2
Session Chair: Wook-Shin Han (POSTECH)
From Walk-Level GNNs to GraphRAG: Towards Knowledge-Enhanced Graph Intelligence
Keynote 2
Session Chair: Wook-Shin Han (POSTECH)
From Walk-Level GNNs to GraphRAG: Towards Knowledge-Enhanced Graph Intelligence
Abstract
Graph neural networks (GNNs) have advanced graph representation learning, yet conventional GCN-derived architectures struggle on heterophilous graphs, where nodes with different labels are more likely to connect. In the first part of this talk, we introduce walk-level GNNs, a new paradigm leveraging fine-grained walk-based features. We present the Label Context Classifier (LCC), which captures higher-order label connectivity via multi-walk embeddings, and WalkFusion, which decomposes GNN message passing into five distinct walk channels (self, forward, backward, sibling, guardians) and adaptively fuses them. These methods achieve notable accuracy gains while improving interpretability.
In the second part, we turn to Graph Retrieval-Augmented Generation (GraphRAG), an emerging approach that integrates knowledge graphs with large language models (LLMs). GraphRAG addresses two key challenges of LLMs: excessive model size and limited domain-specific knowledge. We describe a pipeline that converts natural-language queries into graph queries, performs Prize-Collecting Steiner Tree (PCST)-based retrieval, and embeds the retrieved results into LLM prompts. We conclude by highlighting open research challenges, including efficient graph retrieval, query optimization, and scalable integration of symbolic and neural reasoning.
Bio
Makoto Onizuka is a Professor (Executive Assistants to the President) at Graduate School of Information Science and Technology, Osaka University. He is the leader of Big data engineering Laboratory and conducts research on graph mining algorithms and AI-driven database query optimization techniques. Prior to joining Osaka University, he worked at Nippon Telegraph and Telephone Corporation (NTT) for more than 20 years being served as a distinguished technical member from 2010 to 2014. He also worked as a visiting scholar at the University of Washington from 2000 to 2001.
He developed research prototype systems and some of them were used in production: LiteObject (object-relational main memory database system), XMLToolkit (XML stream engine), CBoC type2 (Common IT Bases over Cloud Computing at NTT), and Grapon (Graph mining techniques).
He serves as Director at the database society of Japan (2024-present), Information Processing Society of Japan (2019-2021), Review Board for PVLDB (2025), Academic committee at Shonan meeting (2016-present), PC co-chair at DASFAA 2024 and KJDB 2024, Workshop co-chair at SIGMOD 2027, Publicity co-chair at MIPR 2021, Workshop co-chair at VLDB 2020, Best demonstration award committee at DASFAA 2010, Best paper award committee at DASFAA 2012 and program committee at international conferences, including VLDB(2024 demonstration track, 2009-2010 industrial track), SIGMOD(2018, 2020, 2027), ICDE(2015 industrial track, 2023 demonstration track), AAAI(2021-2025), NeurIPS(2022-2025), ICLR(2024), IJCAI(2023-2025), ICML(2023-2024), ECML(2022, 2024,2025), CIKM(2017-2018,2020-2023), DASFAA(2010-2016,2020-2023, Senior PC 2025).