The Faculty Early Career Development Program (CAREER) is considered one of the highest recognized forms of achievement with the NSF. As a Foundation-wide movement, the NSF acknowledges and selects early-career faculty who can serve as leaders in research and education, to further advance the organization's mission.   

 The NSF grant provides a significant opportunity to advance this work over the next five years, enabling deeper exploration and real-world integration of their research findings.  

Ziani’s research aims to provide mathematical foundations for long-term fairness in algorithms – particularly those that significantly impact people's lives, such as job placements, loan approvals, and educational opportunities.  

 

Identifying Pipeline Problems  

Motivated by both professional and personal experiences, Ziani is driven to explore fairness beyond surface-level adjustments, to understand and address the long-term impacts of algorithmic decisions on diverse populations.   

An interesting challenge in Ziani's research is creating new mathematical models that accurately reflect the long-term impacts of algorithmic decisions.   

Current models often fail to account for dynamic human responses and the compounding nature of decisions over time.   

“We don't just make one decision right now about a person in real life, we make many, many, many, many decisions about people and we have to think about how all of these decisions are going to interact with each other.”   

Through his work, Ziani wants to emphasize the importance of early intervention in educational pipelines to prevent long-term disparities and the intergenerational effects of an individual’s socioeconomic status.  

To mitigate this, he wants to further explore the value of interdisciplinary collaboration and grounding theoretical work in empirical data.  

  

Connections and Contributions  

By collaborating with education researchers, Ziani will gather and analyze data to better understand when and how to correct disparities through various interdisciplinary approaches.  

Key collaborators in Ziani's research such as Tuba Ketenci, Director of Educational Outreach, and other experts in education and social economics, are essential for understanding the complexities of education pipelines and ensuring the practical applicability of his models.  

“It's really hard to fix those problems when you're intervening so late, like at the university admissions level, after disparities have compounded over years.”  

By developing mathematical and game theory models, Ziani plans to offer more comprehensive and systematic approaches to fairness in AI.   

Ziani hopes his research will push the field of AI fairness to consider broader and longer-term impacts. With acknowledging the risks of developing incorrect models, he recognizes that those barriers could lead to negative real-world outcomes.  

Looking ahead, Ziani aspires to bridge the gap between theoretical and practical applications in AI fairness, with an emphasis on moving away from static modeling of human behavior and to consider the dynamic, long-term effects of continuous decision-making.  

By advancing new mathematical models and collaborating with experts across disciplines, Ziani seeks to address the never-ending complexities of real-world applications and contribute to a more equitable future.  

__________

Author: Camille Carpenter, Communications Manager 

Juba Ziani