TOKYO, June 18, 2025 - Mitsubishi Electric Corporation (TOKYO: 6503) announced today that it has developed a language model tailored for manufacturing processes operating on edge devices. The Maisart^®1-branded AI technology has been pre-trained with data from Mitsubishi Electric’s internal operations, enabling it to support a wide range of applications in specific manufacturing domains. In addition, the model leverages a uniquely developed data-augmentation technique to generate responses optimized for user-specific applications.

The widespread adoption of generative AI is accelerating the use of large language models (LLMs). However, the significant computational and energy costs associated with LLMs are a growing concern. Additionally, there is increasing demand for generative AI solutions that can operate in on-premises² environments due to data privacy and confidential information management requirements.

In response, Mitsubishi Electric has developed a domain-specific language model by training a publicly available Japanese base model with the company’s proprietary data from its own business domains, including factory automation (FA). Using training data generated through the company’s original augmentation techniques enabled effective, task-specific fine-tuning. The resulting model is compact enough to run on limited hardware resources, making it suitable for environments with constrained computing capabilities such as edge devices, as well as for on-premises operations such as call centers that handle sensitive customer information.

The language model was developed as part of the AWS Japan Generative AI Accelerator Program³ provided by Amazon Web Services Japan G.K. AWS Japan provided support for the development, including computing resources such as GPUs⁴ and AWS Trainium⁵ for language-model training, assistance in building distributed training environments, AWS credits, and scientific support from the AWS Generative AI Innovation Center.⁶