HPE’s new solution; HPE Swarm Learning, a distributed machine learning approach, allows users to share what they learn at the edge or in distributed locations without compromising data privacy.

Hewlett Packard Enterprise (NYSE: HPE) has launched HPE Swarm Learning, a breakthrough new AI solution to accelerate insights into problem solving from disease diagnosis to credit card fraud by sharing and combining learning from AI models without compromising to data privacy.

Developed by Hewlett Packard Labs, HPE’s R&D organization, HPE Swarm Learning is the industry’s first privacy-protecting and decentralized machine learning method for edge or distributed sites.1 The solution provides customers with easy integration with AI models through the HPE Swarm API. container. This gives users the opportunity to quickly share what has been learned from the artificial intelligence model with stakeholders inside and outside the organization without disclosing real data.

Justin Hotard, HPE vice president and general manager of HPC and AI, said: “Herdleer offers a powerful approach to solving problems such as anomaly detection that helps improve healthcare, fraud detection and predictive maintenance. HPE makes a meaningful contribution to the herd learning movement through provide an enterprise-grade solution that enables organizations to collaborate, innovate and accelerate AI models while preserving the ethical, data privacy and governance standards of any organization.”

Introducing the new AI approach to securely use insights at the edge

Most AI model training today takes place in a centralized location based on aggregated data sets. However, this approach can be inefficient and costly, as large amounts of data must be moved back to the same source. In addition, this process can be limited by data privacy, data ownership rules and regulations that restrict the sharing and movement of data, potentially leading to inaccurate and biased modeling. By leveraging training models and endpoint insights, companies can make faster decisions and achieve better experiences and results. Moreover, simply passing on what has been learned from one data source institution to another has the potential to unify industries around the world and further develop artificial intelligence with huge commercial and societal implications.

Sharing data externally; data governance presents a challenge to organizations that require regulatory or compliance compliance and need data to stay where it is. Uniquely, HPE Swarm Learning enables organizations to leverage distributed data that augments the training-centric dataset and build machine learning models fairly without violating data governance and privacy rules. HPE Swarm Learning uses blockchain technology to securely engage members, select leaders, and combine model parameters to provide flexibility and security to the swarm network to ensure only the learning is shared, not the data itself. In addition, HPE Swarm Learning helps remove bias by increasing the accuracy of models.

Swarm learning to enable AI for better purposes

HPE Swarm Learning helps different industries collaborate and develop insights:

• hospitals Inferences can be drawn from imaging images, CT and MRI scans, and gene sequencing data that must be transferred from one hospital to another to improve the diagnosis of diseases and other conditions, while preserving patient information.
• Banking and Financial ServicesBy sharing its knowledge of fraud with multiple financial institutions at once, 2 is better positioned to fight the projected more than $400 billion in credit card fraud worldwide over the next decade.
• production facilities, can use predictive maintenance to understand equipment repair needs and respond before equipment failure and service interruptions. Using herd learning, maintenance department managers can gain better insights from sensor data across multiple production sites.

Examples of use cases from early adopters of HPE Swarm Learning include:

University of Aachen working on histopathology to speed up colon cancer diagnosis

A team of cancer researchers from RWTH Aachen University Hospital in Germany conducted a study that could advance the diagnosis of colon cancer by applying artificial intelligence to image processing to predict genetic changes that could cause cells to become cancerous.

Researchers trained AI models using HPE Swarm Learning on three patient groups from Ireland, Germany and the US and validated the performance of predictions from two independent UK datasets using the same swarm learning-based AI models. The results showed that herd learning models used to improve predictions outperform AI models trained on local data alone, taking into account not only patient data but also other locations.

TigerGraph improves anomaly detection to help banks fight credit card fraud

TigerGraph, the graph-based analytics platform, combines HPE Swarm Learning with data analytics running on HPE ProLiant servers with AMD EPYC™ processors to enable efforts to quickly detect unusual activity in credit card transactions. The unified solution improves accuracy when training machine learning models with massive amounts of financial data from multiple geographically dispersed banks and branches.

availabilty

HPE Swarm Learning is available in many countries. For more information, visit: hpe.com/info/swarm-learning

HPE provides complete and turnkey machine learning development solutions

HPE also announced that with its new HPE Machine Learning Development System, it is removing the barriers for businesses to easily build and train machine learning models at scale so they can achieve value faster. The machine learning software platform is an end-to-end solution that integrates computation, accelerators, and networks to develop and train accurate AI models at scale faster.

Source: (BHA) – Beyaz News Agency