With the rapid evolution of generative AI technology, the financial industry is entering a critical transformation phase from digitalization to intelligentization. In response to the highly specialized nature of financial scenarios, stringent compliance requirements, and complex local contexts, APMIC leveraged its integrated advantages in AI model deployment for this project. Based on localized corpora from the Financial Supervisory Commission, the China Academy of Financial Research, and the Department of Digital Development, along with knowledge construction, model fine-tuning, and the distillation of proprietary models, APMIC enhanced the model's ability to understand financial contexts and regulations, thus constructing a comprehensive financial industry model.

APMIC Founder and CEO Po-Han Wu stated, "The financial industry's demand for AI is not only about improving efficiency, but more importantly, about 'trustworthiness' and 'controllability.' We have observed that general-purpose models still face challenges in the financial field, such as insufficient localization and gaps in professional knowledge. Therefore, APMIC is committed to helping the industry build truly implementable AI models through proprietary model construction and data governance capabilities. The importance of this project lies not only in the development of a single technology, but also in the systemic engineering of integrating data and models, which will help strengthen the core competitiveness of Taiwan's financial industry in the AI era."

Currently, APMIC's Traditional Chinese AI model has passed the government's AIEC trustworthiness test, enabling financial institutions to establish a compliant and trustworthy foundation for AI applications. The financial language model will be optimized for local financial contexts, equipping it with the four capabilities of a junior banking professional: "understanding financial knowledge retrieval and certificate learning," "legal interpretation and compliance control," "credit granting and financial analysis," and "investment and market analysis," further improving application accuracy. Furthermore, by supporting private deployment and localized operating architectures, the model can operate in closed environments, effectively ensuring the security of sensitive data while maintaining high performance and flexibility, responding to the financial industry's stringent requirements for information security and risk control.

Looking ahead, financial big data models will not be limited to single application tools, but will gradually evolve into a key foundation supporting innovation in financial services. As application scenarios continue to expand, higher efficiency and accuracy can be achieved through AI technology. APMIC will continue to deepen its presence in enterprise AI platforms and AI architectures, assisting the financial sector, government, and other industries in building proprietary AI capabilities, and propelling Taiwan to establish a competitive technological and application advantage in the global AI development wave.

About APMIC

APMIC (Accelerate Private Machine Intelligence Company) is a leading brand of enterprise-driven AI solutions, founded in Taiwan in 2017. Starting with Natural Language Understanding (NLU) core technology, it has developed "PrivModel," a model fine-tuning and distillation service solution for enterprise model manufacturing, and "PrivStation," a one-stop solution for enterprise privatization. These solutions help enterprises upload data and create their own dedicated ground-based AI brains without requiring any programming. APMIC products are already used by over 1,100 enterprises and government departments across sectors including healthcare, manufacturing, finance, and retail. APMIC is committed to empowering every enterprise to build its own AI brain, making it easy for everyone to participate in the AI era.

For more information, please visit the APMIC official website: https://apmic.ai/