UCSD - University of California - San Diego

06/16/2026 | Press release | Distributed by Public on 06/16/2026 09:53

UC San Diego Researchers Win Best Paper Award for New Approach to Connecting Complex Data Systems

Published Date

June 16, 2026

Article Content

Researchers at the University of California San Diego have developed a new system designed to solve one of modern computing's growing challenges: how to efficiently analyze data spread across multiple types of databases.

The team's paper, MICRO: A Lightweight Middleware for Optimizing Cross-Store Cross-Model Graph-Relation Joins, won the Best Paper Award at the 2026 IEEE International Conference on Data Engineering (ICDE), one of the field's leading conferences.

Today, organizations rarely store all of their data in a single system. Traditional relational databases manage structured information such as financial transactions and medical records, while graph databases specialize in mapping relationships between entities - an increasingly important capability for applications such as fraud detection, cybersecurity, recommendation engines and scientific knowledge graphs.

But querying data across both platforms is often slow, expensive and technically difficult, especially when a query requires data movement across platforms.

"Our work addresses a problem affecting industries ranging from healthcare and finance to scientific computing and artificial intelligence," said lead author Xiuwen Zheng, a doctoral graduate from the UC San Diego Jacobs School of Engineering Computer Science and Engineering (CSE) Department.

To address that challenge, Zheng and her collaborators developed MICRO, a lightweight middleware platform that enables relational and graph databases to work together more efficiently without requiring organizations to rebuild existing infrastructure.

Xiuwen Zheng started her research at the San Diego Supercomputer Center (SDSC) as a CSE M.S. student and then continued into the Ph.D. program at CSE. She was co-advised by Amarnath Gupta of SDSC and Arun Kumar of CSE and Halıcıoğlu Data Science Institute (HDSI), which is part of the Halıcıoğlu School of Data Science and Computing.

"Instead of forcing all data into a single platform, MICRO intelligently coordinates how queries run across multiple systems," Zheng said. "The middleware analyzes workloads, determines where operations should execute and optimizes how data moves between graph and relational stores."

The result is faster, more efficient analytics across increasingly complex data environments - an advance with significant implications for high performance computing, AI and data-intensive scientific research.

Modern scientific workflows routinely combine diverse forms of information, including simulation outputs, sensor streams, metadata catalogs and relationship-based knowledge graphs. Fields such as fusion energy research and drug discovery all depend on connecting structured datasets with complex networks of relationships.

AI systems face similar challenges. Large language models, recommendation engines and autonomous agents increasingly rely on interconnected data ecosystems rather than isolated databases. As data volumes continue to grow, efficiently linking these specialized systems has become a major bottleneck.

Rather than replacing existing databases, MICRO is designed to improve interoperability across heterogeneous and distributed data systems. The approach reflects a broader shift in modern computing, where organizations increasingly rely on specialized tools working together instead of one-size-fits-all platforms.

"By reducing the complexity of cross-system analytics, our goal is for systems like MICRO to help scientists and organizations spend less time managing infrastructure and more time extracting insights from data," Zheng said.

MICRO is part of a longer research arc at UC San Diego known as Project AWESOME to help domain scientists analyze large multi-model social media datasets and later extended to applications in cybersecurity and beyond. The path to ICDE 2026 was not a straight line: the broader line of work faced multiple close rejections at major database venues before this paper was recognized.

"Xiuwen soldiered on, resolutely believing in the project vision, demonstrating technical creativity, and delivering on its potential," said CSE and HDSI Associate Professor Arun Kumar about the work's trajectory. "Her tale of perseverance is one I hope is helpful to students and researchers everywhere. This work also exemplifies the interdisciplinary research ethos across SDSC, HDSI and CSE. I am excited to see more such fruitful interdisciplinary collaborations across campus spanning engineering, the sciences, and the new Halıcıoğlu School of Data Science and Computing."

SDSC Research Scientist Amarnath Gupta, who leads Project AWESOME, added that collaborations with domain scientists - including Professor Molly Roberts in Political Science and HDSI, whose group works with large-scale social media data - were instrumental in surfacing the cross-store, cross-model challenges that MICRO ultimately addresses. The project is supported by the National Science Foundation.

Xiuwen Zheng received her Ph.D. and M.S. from UC San Diego Jacobs School of Engineering CSE Department in 2025 and 2019, respectively. Her Ph.D. work was co-advised by her publication co-authors Amarnath Gupta of SDSC and Arun Kumar of CSE and HDSI.

Learn more about research and education at UC San Diego in: Artificial Intelligence

UCSD - University of California - San Diego published this content on June 16, 2026, and is solely responsible for the information contained herein. Distributed via Public Technologies (PUBT), unedited and unaltered, on June 16, 2026 at 15:53 UTC. If you believe the information included in the content is inaccurate or outdated and requires editing or removal, please contact us at [email protected]