Intel accelerates AI development with AI reference kit

The optimized AI reference kit is designed to help developers and data scientists achieve innovation "acceleration".

Latest news: Thanks to Intel's long-term cooperation with Accenture, Intel has now launched a set of 34 open source AI reference kits to help developers and data scientists deploy artificial intelligence (AI) faster and easier. Each kit includes AI-optimized model code, training data, machine learning pipelines, libraries, and oneAPI components, allowing enterprises to flexibly apply them in local, cloud, and edge environments using different architectures.

"The Intel AI Reference Kit provides an easy, efficient, and cost-effective way for millions of developers and data scientists to develop and expand AI applications in health and life sciences, financial services, manufacturing, retail, and many other industries. Intel is committed to Through its broad product portfolio - AI-accelerated processors and systems, and its investment in the open AI software ecosystem, the reference kit uses a rich set of components from Intel's AI software product portfolio to build a future where AI is within reach. And built on the open, standards-based oneAPI multi-architecture programming model.”

——Dr. Wei Li, Vice President and General Manager of Artificial Intelligence and Analytics at Intel

Why it matters: The Intel AI Reference Kit is built using oneAPI’s open, standards-based heterogeneous programming model and components of Intel’s end-to-end AI software portfolio, including the Intel® AI Analysis Toolkit and the Intel® Distribution OpenVINO™ Toolkit , allowing AI developers to simplify the process of writing applications using AI, enhance existing intelligent solutions, and accelerate deployment. The results prove that compared with the traditional model development process, the Intel AI Reference Kit brings significant performance improvements while making the workflow more time-saving and efficient.

Preconfigured suites simplify AI development across industries such as consumer products, energy and utilities, financial services, health and life sciences, manufacturing, retail and telecommunications. Some examples of how it benefits various industries include:

• Using the AI ​​reference kit designed for enterprise conversational AI chatbot interaction settings, users can increase inference speed by up to 45% in batch mode through oneAPI optimization.

• An AI reference kit designed specifically for visual quality control inspection automation in the life sciences field. After oneAPI optimization, the training speed of visual defect detection is increased by up to 20%, and the inference speed is increased by up to 55%.

• Improves prediction accuracy by up to 25% for an AI reference suite used by developers to predict the health of utility assets and improve service reliability.

AI reference kits can reduce solution time from weeks to days, helping data scientists and developers train models faster and at lower cost, overcoming the limitations of proprietary environments. Driven by oneAPI, AI tools and optimizations can maximize the portability of open accelerated computing applications.

John Giubileo, managing director at Accenture, said: "Working with Intel to develop AI reference kits for the open source community enables our customers to run AI workloads in a more efficient way. These kits built on oneAPI are designed to provide developers with convenient and efficient AI solutions, thereby reducing project complexity and shortening their deployment time in various industries.”

Future Outlook: Intel will continue to update the reference kit based on community feedback and contributions. Including visual quality inspection, enterprise conversational AI chatbot setup, predictive asset health analysis, medical imaging diagnosis, document automation, and AI structured data generation, etc.