Zhu harnessing interpretable neuro-symbolic learning for reliable ranking
2025-09-03
(Press-News.org)
Ziwei Zhu, Assistant Professor, Computer Science, College of Engineering and Computing (CEC), received funding for the project: “III: Small: Harnessing Interpretable Neuro-Symbolic Learning for Reliable Ranking.”
Zhu aims to forge novel neuro-symbolic learning to rank models to deliver interpretable, balanced, and robust rankings.
In this research, Zhu will introduce a neural network that transparently elucidates the entire inference process in ranking and presents a logic AutoEncoder, a type of artificial neural network for interpretable collaborative filtering rankings. He will also propose a neuro-symbolic representation learning technique to enhance model expressiveness and accuracy.
Second, he will utilize the inherent transparency in neuro-symbolic ranking algorithms to effectively enhance ranking balance for both users and items.
Further, he will propose novel strategies for leveraging and augmenting interpretable reasoning process of neuro-symbolic learning to fortify robustness against shortcut features, exposure differences, and data poisoning attacks.
Additionally, Zhu will apply proposed research innovations in two real ranking systems, organ transplantation management and research paper recommendation. He will do this to demonstrate the effectiveness of proposed technologies in real-world contexts and to ensure the proposed project has sufficient impact.
Zhu received $500,000 from the National Science Foundation for this research. Funding began in July 2025 and will end in late June 2028.
###
ABOUT GEORGE MASON UNIVERSITY
George Mason University is Virginia’s largest public research university. Located near Washington, D.C., Mason enrolls more than 40,000 students from 130 countries and all 50 states. Mason has grown rapidly over the past half-century and is recognized for its innovation and entrepreneurship, remarkable diversity, and commitment to accessibility. In 2023, the university launched Mason Now: Power the Possible, a one-billion-dollar comprehensive campaign to support student success, research, innovation, community, and stewardship. Learn more at gmu.edu.
END
ELSE PRESS RELEASES FROM THIS DATE:
2025-09-03
Weiwen Jiang, Assistant Professor, Electrical and Computer Engineering, College of Engineering and Computing (CEC), and Lei Yang, Assistant Professor, Information Sciences and Technology, CEC, received funding for a workshop on quantum system stability and reproducibility.
Jiang and Yang will use this funding to support students and researchers who participate in StableQ 2025. The event will bring together experts from academia, industry, and national laboratories to report state-of-the-art developments, exchange ideas and practices, and foster ...
2025-09-03
Li Studying Quantum Algorithms
Fei Li, Associate Professor, Computer Science, College of Engineering and Computing (CEC), received funding for the project: “Quantum Algorithms for High-Performance Analysis of Single-Cell Omics Data and Explainable Drug Discovery.”
This project leverages quantum computing to develop innovative, explainable methods for drug target discovery by integrating biological omics data—such as single-cell RNA sequencing (scRNA-seq) from disease tissue samples—with ex vivo drug ...
2025-09-03
Long-term use of benzodiazepine and related drugs has been linked to poorer quality sleep in older adults with insomnia. While they are cheap, commonly prescribed and widely available in Quebec, these drugs are also addictive: quitting cold turkey can lead to serious withdrawal symptoms, including increased risk of falls, hospitalization and cognitive issues.
A new study led by Concordia researchers published in the journal Sleep shows that use of benzodiazepine and benzodiazepine receptor agonists (BZRA) can also affect brain rhythms during sleep that are important for memory and cognitive health in adults aged between 55 and 80.
“These drugs don’t ...
2025-09-03
Key takeaways:
A PICALM gene variant disrupts microglia, the brain’s cleanup cells, reducing their ability to clear harmful proteins and lipids.
This dysfunction leads to lipid droplet buildup in microglia, weakening their protective role and raising Alzheimer’s risk.
The findings provide a roadmap for drug development targeting genetic risk factors in late-onset Alzheimer’s disease.
TAMPA, Fla. (Sept. 3, 2025) – A Nature study published today and led by researchers at the University of South Florida’s USF Health Byrd ...
2025-09-03
Confiscating personal belongings during government-led dismantling of tent cities in Vancouver inflicts immediate harm and further destabilizes people already struggling to meet their basic needs, according to a new study from Simon Fraser University.
Published in the journal Public Health, the study found nearly one in four people experiencing homelessness reported having their personal belongings confiscated by city workers between 2021 and 2023. These confiscations—often part of street sweeps to remove tent cities—were significantly associated with non-fatal overdoses, violent victimization, and barriers to accessing essential services.
“Our data captures ...
2025-09-03
FOR IMMEDIATE RELEASE, September 3, 2025
Contact: Gina DiGravio, 617-358-7838, ginad@bu.edu
New Study Seeks to Understand the Links between Social Drivers of Health by Investigating Cardiovascular Health in Young Adults
“Detailing childhood social determinants helps target those factors that drive cardiovascular disease in adulthood”
(Boston)—Cardiovascular disease (CVD) remains the leading cause of death and disability for adults in the U.S. Recent projections from the American Heart Association suggest that by 2050, more than 45 million American adults will have clinical CVD and more than 184 million will ...
2025-09-03
(Santa Barbara, Calif.) — Using reprogrammed biocatalysts, researchers are pushing the boundaries of enzymatic synthesis with a method that opens the door to a diverse array of valuable compounds. Reporting in the journal Science, UC Santa Barbara chemistry professor Yang Yang and collaborators detail an enzymatic multicomponent reaction, resulting in six distinct molecular scaffolds, many of which were not previously accessible by other chemical or biological methods.
“The ability to generate novelty and molecular diversity is particularly important to medicinal chemistry,” Yang said. “For a long time, biocatalysis ...
2025-09-03
Researchers from University of California San Diego have found that Delta-8 tetrahydrocannabinol (delta-8 THC), a psychoactive compound often sold as a legal alternative to marijuana, is most commonly used in states where marijuana use remains illegal and delta-8 THC sales are unregulated. The findings, published in the American Journal of Preventive Medicine, highlight how gaps in cannabis policy may be inadvertently steering people toward less-regulated substances and have allowed manufacturers to evade restrictions ...
2025-09-03
OKLAHOMA CITY – A University of Oklahoma study published Sept. 3 in JAMA Surgery reports that acute normovolemic hemodilution (ANH) – a blood-saving method in which a patient’s blood is collected before going on heart-lung bypass and reinfused near the end of cardiac surgery – remains underused in the United States at 14.7%. Yet the study found that ANH lowered the likelihood of a transfusion by 27%, a decrease in blood use that could cut costs substantially while still protecting patient safety and outcomes.
Global demand for cardiac surgery ...
2025-09-03
Using the first complete dataset of more than 415 million buildings across 50 countries in sub-Saharan Africa, researchers at the University of Chicago created an unprecedented approach to urban development, down to each street block.
The new analysis, published this week in Nature, pinpoints where rapidly developing nations lack “last mile” infrastructure and access to public services. It uses high-resolution data to measure street access to each building across the subcontinent, showing ...
LAST 30 PRESS RELEASES:
[Press-News.org] Zhu harnessing interpretable neuro-symbolic learning for reliable ranking
