Intel Partners with Accenture to Launch Open Source Pre-Trained AI "Reference Kits"

Intel, in partnership with Accenture, has announced the launch of new open source "reference kits" for artificial intelligence (AI) project development, originally developed at the company's computer vision arm — and, it claims, designed to make AI technology more accessible.

"Innovation thrives in an open, democratized environment. The Intel accelerated open AI software ecosystem including optimized popular frameworks and Intel’s AI tools are built on the foundation of an open, standards-based, unified oneAPI programming model," Intel's Wei Li, PhD, explains. "These reference kits, built with components of Intel’s end-to-end AI software portfolio, will enable millions of developers and data scientists to introduce AI quickly and easily into their applications or boost their existing intelligent solutions."

The company is opening with four key reference kit releases: Utility asset health, designed to model the health of utility poles through 34 attributes; visual quality control, an optimized computer vision system the company claims offers 95 percent accuracy in pinpointing pharmaceutical pill defects; a customer chatbot, built around BERT and PyTorch and offering a claimed 45 percent boost to inferencing performance compared to its unoptimized predecessor; and intelligent document indexing, designed to offer 65 percent accuracy in reviewing and sorting business documents.

Each of the reference kits is designed to solve a real-world problem and comes pre-trained, but is also being made available under an open source license — meaning each can be used as the starting point for a solution to a related problem.

The four kits have been published to the OneAPI GitHub repository under the permissive BSD 3-Clause license, with more information and a link to each kit's individual repository available on Intel's website.