Skip to main content
Exploring what’s next in tech – Insights, information, and ideas for today’s IT and business leaders

Swarm learning: Driving advances both practical and profound

This episode of On the Goh highlights how swarm learning has the potential to solve some big problems, from helping hospitals diagnose disease to tackling credit card fraud.

Data is a powerful currency in today's global economy, enabling advances such as artificial intelligence and machine learning. But for enterprises to maximize the full value of their data, they must be able to share it not only within their own organizations but also with others.

"That's where challenges arise," says Dr. Eng Lim Goh, senior vice president and chief technology officer for AI at Hewlett Packard Enterprise.

Please read: Getting value from your data shouldn't be this hard

For regulatory and proprietary reasons, data is often siloed within organizations, he says, preventing them from bringing together multiple data sources for greater insights. But a new approach to sharing learnings without sharing the data itself is changing that: swarm learning. With swarm learning, "there are massive opportunities for organizations to improve customer experience, create more intuitive products, and even solve some of the world's most complex and pressing problems," Goh says.

How swarm learning works

As he explains in this On the Goh episode, "swarm learning is a decentralized machine learning solution that uses edge computing and is built on blockchain technology for peer-to-peer collaboration." Using a private-permission blockchain, only insights—not the data itself—are shared, which ensures data security and privacy while allowing all to benefit from the collective learnings, he says.

Please read: How swarm learning provides data insights while protecting data sovereignty

For example, in the case of a hospital that is building a machine learning model for diagnostics, swarm learning would enable it to combine its protected patient information—imaging records, CT and MRI scans, gene expression data, and more—with that of hospitals in other regions and generate insights that can then be shared among all the participants. Such a model not only provides greater intelligence but prevents biases based on a particular hospital's demographic.

"Imagine the value to humankind if we united the hospitals of the world," Goh says.

The answer to problems today—and tomorrow

Another example can be found in the finance industry, where swarm learning promises a solution to problems like credit card fraud, which is projected to surpass $35 billion globally by 2025.

Most financial companies use their own data to train machine learning models to detect fraud, but they have "different strengths and weaknesses," Goh says. Companies could better protect their businesses and customers by sharing learnings from fraud data without revealing proprietary customer information.

"These are just two examples of how swarm learning has the potential to change our lives in ways that are both practical and profound," he says.

Watch other On the Goh episodes.

This article/content was written by the individual writer identified and does not necessarily reflect the view of Hewlett Packard Enterprise Company.