IBM Federated Learning and Data Fabric: Preparing for the Next Pandemic

As we’ve recently discovered, pandemics are still a thing—and a thing that is likely to recur at an increasing cadence either due to variants of current viruses or the emergence of new ones. IBM was at the heart of the effort to find remedies early on and ran into significant problems with data access for medical repositories and assuring that regulations regarding privacy were met. IBM is driving two efforts that could significantly not only speed how companies can more rapidly respond to tactical and strategic threats but more rapidly identify remedies for these future viral events.

These two technologies are Federated Learning and IBM Data Fabric

Let’s talk about why these two technologies are potentially a game changer when mitigating the next pandemic.

IBM Federated Learning

The big issue, as noted above, when trying to locate existing medications that might show promise in mitigating or curing the COVID-19 pandemic wasn’t the lack of data, it was that there not only was too much data, but it was also in databases all over the world in inconsistent formats and protected by laws or practices that made sharing the data nearly impossible.

I expect we’ll eventually find that these siloed data practices resulted in millions of avoidable deaths. And data access, if anything, appears to be getting worse because, for some screwy reason, the justifications for keeping the data private seem to exceed the justification for saving lives.

What IBM Federated Learning does is use a process where the data is analyzed in place. Only the analysis results are forwarded to an aggregator obfuscating any of the individual details of the patents involved or other projects that a research hospital might be undertaking. Federated Learning potentially protects both the patients’ privacy and most of the competitive advantage of the research hospital.

IBM Data Fabric

IBM Data Fabric is a common trusted data foundation delivering business-ready data at extremely high speeds. It identifies data sources and dynamically orchestrates those sources to provide the selective information needed to satisfy the query. This solution is agnostic to deployment platforms, data process, data use, geographical locations, and architectural approach. It has built-in compliance with governance, security, and regulations surrounding the data, and it has proven to yield cost and operational efficiencies by eliminating the need for independent tools.

In short, added to IBM Federated Learning, IBM’s Data Fabric provides what a supercharger might provide on an internal combustion engine, a massive speed boost. And it appears well designed to deal with the exact problems identified during the effort to find a COVID-19 remedy. It can operate across securely protected and regulated medical databases while complying with the local laws surrounding data privacy and protecting the intellectual property of the database owners unrelated to the specific antiviral work.

Together

Together IBM Federated Learning and IBM Data Fabric create a foundation that—if the world’s medical data repository owners allow it—would provide a far more rapid ability to identify remedies for future viruses in a fraction of the time we saw with the current impressively fast effort with the increased possibility of avoiding many of the objections surrounding the current efforts. For instance, provisional approvals of antiviral protocols and other remedies could be replaced by unqualified approvals giving potential patients more confidence in the result.

Wrapping Up: Layering in AI

With IBM Federated Learning and IBM Data Fabric, the potential to develop remedies to future illnesses and pandemics far more rapidly becomes possible. But, IBM is also an AI company, and if you layer on IBM Watson, suddenly you have the ability for doctors—or maybe individuals—to be able to get faster remedies for an increasing variety of hard to treat illnesses and better remedies that are more customized for common illnesses due to having more data to analyze and an increasingly smarter AI able to provide the result in natural language.

Thanks to tools like these, we may more quickly overcome future pandemics and even avoid a growing number of them. And these same tools can be used to solve an increasing number of business problems and even create new products like auto insurance that adjusts the premiums for how and how often you drive, not to mention future health insurance that is priced based on your overall health and behavior.

Our data-driven future is coming, and IBM Watson, IBM Federated Learning, and IBM Data Fabric, if used correctly, could make our future far brighter than it otherwise would be.