Shared Artificial Intelligence without data sharing: Data privacy in a hub and spoke architecture for modeling and managing global flows
Globalization and technological innovation have dramatically accelerated the global flows of capital, labor, cargo, people, goods, services, ideas, images, data, and electrons. In order to direct global flows toward shared, constructive, and legal ends, nation states and the private sector must gain visibility over them, and then distinguish, in real time, between the good and the bad, the high risk and low risk, the legal and illegal. Recent breakthroughs in privacy- and sovereignty-preserving data collaboration and machine learning technologies facilitate knowledge-graph construction and adaptive insight generation at scale — illuminating transnational networks, characterizing and risk-manage global flows, and empowering institutions that have been painfully diminished by the rise of transnational forces.
Peter Swartz demonstrates a “hub-and-spoke” model that can facilitate secure and private accessing, linking, and generating insight from data describing global flows. These hub-and-spoke infrastructures benefit from advances in software containerization, and deployment across customer-specific, virtual private clouds. Machine learning training and productionalization can thus be accomplished while meeting strict requirements around client data privacy, security, and sovereignty. This general architecture is applicable to many industries.
Peter Swartz is cofounder and CTO at Altana Trade, an artificial intelligence partnership unlocking the power of global economic data to make trade safer, more efficient, and more profitable. Previously, Peter led the data science team at Panjiva—ranked as one of Fast Company’s most innovative companies in 2018 and acquired by S&P Global.
For conference registration information and customer service
For more information on community discounts and trade opportunities with O’Reilly conferences
For information on exhibiting or sponsoring a conference
For media/analyst press inquires