Robotic hand presses an on-screen button

Addressing the data challenges of industrial AI

Data challenges are a major concern for upper-level management in the manufacturing sector currently implementing industrial AI, according to a recent MIT Technology Review Insights article, Taking AI to the next level in manufacturing.

An in-depth survey of over 300 top executives at leading aerospace, automotive, chemical, high tech, and machinery manufacturing companies across five continents find that scaling up industrial AI implementations can stall without a solid foundation of high-quality, accessible data.

The main challenges

Participants in the survey cited inadequate data quality and weak data integration as the main data challenges.  “Poor data quality results from a variety of factors,” the report says.  “Errors in data entry, missing data points, inoperative sensors in plant equipment, and siloed data trapped in legacy systems are just some of the more common ones.”

The report goes on to explain that data silos are the result of poor data integration, and are a key obstacle to scaling AI use cases up to the enterprise-wide level.

“Different parts of plants have different data systems associated with them, especially if they were built decades ago,” said Gunaranjan Chaudhry, director of data science at SymphonyAI Industrial, quoted in the report. “The data is in vastly different places and difficult to bring together to build good AI models.  Even new facilities were designed before people realized that having all this data in one place allows them to do a lot of things with it.”

Integrating the data

The challenge, then, is collecting this data from different sources, often with different data protocols, integrating it into a single unified namespace, and then sending it securely from the OT (operations) network to an AI system, typically residing on a cloud server.

This is exactly what Skkynet technology does, and has been doing for years.  Our DataHub software allows users to connect, concentrate, integrate, and redistribute their data among sources and users across the enterprise.  And our unique approach to industrial AI provides the most secure way to connect an industrial network to a cloud server, through a DMZ.

We understand that there are many challenges to building a functional, profitable industrial AI solution.  Data challenges are just part of a larger picture.  But when it comes to integrating diverse data sets and connecting them securely to a cloud-based AI system, Skkynet has viable solutions.