Qiunan Zhang
Alabama State University, AL·Computer Science
Online reviews and social media posts carry hidden information about the people who wrote them, such as education, personal preferences, location or language. These hidden variables, called confounding factors, shape what people write.
- Fields
Computer Science
- Topics
machine learning
confounding factors
confounding bias
- Grants
- 1 known grant, through Sep 2028
What they work on
- Build machine learning models that identify key confounding factors, such as education, location or language, hidden in textual and non-textual user-generated data like online reviews.
- Improve confounder-detection models through iterative learning that brings in multi-source and multimodal data, such as social media posts combined with reviews.
- Reduce confounding errors and strengthen model reliability using multi-view learning within a contrastive learning framework, which learns by comparing similar and dissimilar examples.
More about the research
They can mislead machine learning models that use this content to support decisions. A model may then work poorly on new data or in real settings such as healthcare or e-commerce. This project aims to find and reduce these confounding errors so that decision-support systems are more accurate and robust.
Why it matters
Decision systems in healthcare and online commerce rely on data that carry hidden biases from user backgrounds. Reducing these confounding errors could make machine learning recommendations more reliable when they are applied to new users and settings. Industry partners such as FedEx may help move the methods into real decision-support applications.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
Closest labs
Kenneth Joseph
University at Buffalo, NY·Computer Science·Early-career
Develop mixed methods combining graph-based and natural-language machine learning with qualitative analysis to identify dominant beliefs about social issues in large datasets from X and Parler.
Faez Ahmed
Massachusetts Institute of Technology, MA·Engineering·Early-career
Develop diffusion models, generative AI that builds outputs by removing noise step by step, that learn from the full history of design optimization runs rather than only final optimized designs, to generate more precise engineering designs.
Meng Jiang
University of Notre Dame, IN·Computer Science·Early-career
Develop and compare natural language generation models, one data-driven and one based on theories of abduction, for comparative summarization and argument generation in scientific domains.
The school
Alabama State University: 2 labs on PI Pages, $4M a year on research, 7 research doctorates a year.
Reach out
A start for your email. Fill in the brackets.