沪语语音大模型
Search documents
首创沪语交互AI应用 上海银行携手财跃星辰、阶跃星辰共创智慧养老金融服务新范式
财联社· 2025-11-04 05:09
Core Viewpoint - The strategic collaboration between Shanghai Bank, Caiyue Xingchen, and Jiyue Xingchen marks the launch of China's first comprehensive financial model system focusing on elderly finance and dialect intelligence, creating a new paradigm of smart elderly financial services through "technology innovation + scenarios" [4]. Group 1: Technological Breakthroughs - The collaboration introduces a pioneering Shanghai dialect full-link interaction solution, addressing the unique needs of elderly clients who prefer using dialects [5]. - Unlike common banking models, this initiative employs a multimodal large model tailored for banking scenarios, enhancing understanding and processing capabilities by integrating text, voice, image, and video data [5]. - The Shanghai dialect voice model enables a complete interaction chain from voice wake-up to feedback, allowing clients to conduct banking business in their dialect [5]. Group 2: Service Innovations - The AI-driven mobile banking service allows clients to complete over ten high-frequency services, such as transfers and pension management, through text or voice commands, enhancing user experience for elderly clients [6]. - The service integrates non-financial scenarios like dining and travel, creating a "life + finance" one-stop service, positioning the mobile banking app as a daily life assistant [6]. - Future developments will include AI assistants and intelligent investment advisors that predict client needs and provide personalized content, achieving a "one-to-one" service model [7]. Group 3: Strategic Layout - The collaboration aims to establish the industry's first integrated financial large model service system, addressing data quality challenges in AI application within the financial sector [8]. - Caiyue Xingchen, a company focused on financial large models, leverages high-quality financial data to empower the implementation of AI in the financial industry [8]. - The self-developed multimodal Caiyue large model, trained on trillions of financial data, excels in financial knowledge Q&A and financial chart understanding, showcasing superior capabilities compared to similar products [8].