eIQ, a comprehensive machine learning toolkit
Supports the entire NXP MCU product line
On the 18th, NXP Semiconductors unveiled the eIQ edge intelligence software environment and solutions tailored for professional applications to enable edge node developers to leverage machine learning.

eIQ is used in a wide range of industrial, IoT, and automotive applications. It includes the tools necessary to configure and optimize machine learning models trained for the cloud to run efficiently on resource-constrained edge devices. This production-ready turnkey solution is specialized for voice, vision, and anomaly detection applications. While previously, professionally adopting machine learning required massive investment, NXP supports tens of thousands of customers who wish to integrate machine learning capabilities into their products without this burden.

eIQ supports NXP's entire line of MCUs and application processors. Keeping pace with the changing landscape of machine learning, eIQ provides data collection and curation tools, as well as various neural network (NN) frameworks such as TensorFlow Lite, Caffe2, CNTK, and Arm NN through continuous expansion. In addition, it includes model transformation for inference engines, support for new NN compilers such as GLOW or XLA, existing machine learning algorithms, and tools for model deployment for heterogeneous processing on NXP embedded processors.

NXP recently introduced a software infrastructure called EdgeScale to enable machine learning applications, integrating data collection, curation, and processing at the edge. This allows for the seamless integration of cloud-based AI and machine learning services. Additionally, cloud training models and inference engines can be deployed on all NXP devices, ranging from low-cost MCUs to high-performance i.MX and Layerscape application processors.

NXP has also launched a turnkey solution for the local execution of edge-based learning, vision, speech, and anomaly detection models built in the eIQ environment. This system-level solution enables customers to implement differentiated capabilities and provides the hardware and software necessary to build fully functional applications. As a modular solution, product capabilities can be easily expanded via a plug-in method.

“We are well aware that processing at edge nodes is what drives real-world machine learning adoption by customers,” said Geoff Lees, Senior Vice President and General Manager of Microcontrollers at NXP. “NXP builds scalable machine learning solutions and eIQ tools to help customers more easily utilize artificial intelligence capabilities from the cloud to the edge.”