Of course, there are risks to a multifunctional LTM approach. Failures in widely deployed models can impact the entire system. This goes some way to explaining Mastercard’s strategy, at least for now, to apply its technology in parallel with existing detection systems.
Mastercard wants to increase the scale and overall sophistication of the data used in this model. We also plan API access and SDKs to enable internal teams to build new applications.
This blog post highlights the data responsibilities that LTM has and addresses privacy and transparency, model explainability, and auditability. In addition to the data handling involved in LTM operations, we expect regulatory oversight of systems that influence credit decisions and fraud outcomes.
Highly structured data, as opposed to text or images, is at the core of LTM. Large-scale tabular models could be the beginning of a new generation of AI systems in core banking and payments infrastructure. Evidence to date is limited to vendor reports, so performance claims should not necessarily be considered conclusive.
Robustness under adversarial conditions, long-term post-training costs, and regulatory acceptance are all issues on which tabular models can establish or thrive. These factors will determine the pace and scope of implementation, but this is where Mastercard is betting right now.
(Image source: “Oversight” from the official U.S. Marine Corps page is licensed under CC BY-NC 2.0.)
Want to learn more about AI and big data from industry leaders? Check out the AI & Big Data Expos in Amsterdam, California, and London. This comprehensive event is part of TechEx and co-located with other major technology events. Click here for more information.
AI News is brought to you by TechForge Media. Learn about other upcoming enterprise technology events and webinars.
