Skip to main content
SearchLoginLogin or Signup

Center for TAIMing AI

The Center for TAIMing AI is a pilot center at the University of North Carolina at Charlotte

Published onNov 16, 2023
Center for TAIMing AI

The pilot Center for TAIMing AI aims to develop the methods to understand the limits and weaknesses of Machine Learning (ML) and Artificial Intelligence (AI) algorithms such that their adoption and use betters our society. This emerging research area is grounded in the Charlotte area banks’ strength in model risk management (MRM). The Center will focus on three main themes of Trustworthy AI through MRM (TAIM), mainly 1) extending the uncertainty quantification framework to establish clear demarcations for when an algorithm functions reliably and when it encounters limitations; 2) establish a hierarchical control framework that accounts for and ensures robust, resilient, and reliable implementation of algorithms; and 3) formulate a comprehensive framework managing the use of algorithms within our society that is based on 1) and 2).

The overarching goal of the center is to be an international research hub in the emerging area of identification and management of risks associated with the adoption of algorithms. Entities adopting algorithms and predictive models have often driven decisions without understanding the underlying risk and uncertainty to customers and our society. Thus, there is a pressing need to develop the rigorous scientific framework to ensure algorithms are trusted given the natural uncertainty and biases in models. This nascent subfield requires interactions of several research communities including machine learning, data science, applied mathematics, statistics, computer science, humanities, social sciences, risk management, ethics, defense, and health. Being grounded through partnerships with the Charlotte banking sector brings a knowledge base of managing the risk of adopting algorithms in practice as these banks have rapidly developed and adopted MRM practices over the last 15 years (post the 2008 Great Recession and regulatory oversight through the Federal Reserve's SR 11-7).

The center will deepen the intellectual foundations of TAIM methods, open tools, and processes. The center will focus on creating new knowledge, algorithms, open software, and risk and regulatory frameworks in this emerging field of research through partnerships with academia, industry, government (as demonstrated through our team and letters of support), and community stakeholders. The center will offer workshops, short courses, industry internships, researchers in residence, as well as faculty sabbaticals in industry. One of the goals of this center is to extend the intellectual foundation of these practices across a broader set of modeling techniques and practices where entities are adopting algorithms. This center will be the intellectual center for not only leading-edge techniques, but also sound adoption practices for industry and governmental entities.

Comments
0
comment
No comments here
Why not start the discussion?