NXP Semiconductors is responding to the growing interest in artificial intelligence (AI) that operates on edge devices like microcontrollers (MCUs) and microprocessors (MPUs). They’ve rolled out new tools to help developers implement AI on these edge processors more easily and efficiently.
NXP emphasizes that running AI at the edge comes with perks like faster responses, lower energy use, and better privacy for data. Their upgraded eIQ Toolkit is designed to speed up this process, offering developers access to various AI models – from generative AI to time series and vision models. Plus, these models can be deployed on a broader range of edge processors.
One major addition is GenAI Flow with retrieval augmented generation (RAG). This feature allows for fine-tuning models based on specific domain knowledge and private data without needing to fully retrain the original model, making generative AI more accessible for edge devices. They’ve also revamped the eIQ Time Series Studio and their AI development software to streamline the AI deployment process on both small MCUs and larger MPUs.
GenAI Flow is essential for developing large language models (LLMs) that support generative AI applications. It works well with MPUs like NXP’s i.MX series, focusing on employing contextual data to train LLMs effectively. Imagine an appliance with an LLM tailored to its user manual that can chat with users about its features and maintenance – that’s the goal.
The eIQ Time Series Studio is all about making it easier and quicker to develop time series-based AI models. Applications range from detecting water leaks to monitoring temperature and predictive maintenance for machinery. It even helps with energy demand forecasting and optimizing building temperature control.
This studio offers an automated workflow that simplifies the creation and deployment of time series machine learning models on MCU-class devices, including the MCX and i.MX RT series from NXP.
Charles Dachs, NXP’s senior vice president and general manager of industrial and IoT, highlighted AI’s potential to create a more anticipatory and automated world that fits users’ needs. NXP aims to provide developers with tools that suit both small-scale AI models on MCUs like the MCX series and larger generative AI models on higher-end processors like the i.MX 95. NXP is paving the way for practical edge AI solutions that appeal to various markets.