To delve into the cutting-edge concepts of artificial intelligence development in the era of digital intelligence, keenly capture market innovation trends, and fully stimulate industry's innovative vitality, on April 17, 2024, the IT Market Annual Conference AI Forum was held in Beijing. Lin Songtao, Vice President of TRS, was invited to attend and deliver a keynote speech on "Challenges and Practices of Implementing Industry-scale LLMs" and participate in a high-level dialogue. Meanwhile, TuoerSi was honored to be listed in the Top 20 Leading Chinese LLM Enterprises in 2024 and Top 100 Chinese Artificial Intelligence rankings.
This AI forum was organized by CCID Consulting, with numerous experts, scholars, corporate leaders, and industry colleagues gathering together to collectively explore the development path of general artificial intelligence led by cutting-edge technologies such as large language models and generative AI.
In his speech, Vice President Lin Songtao stated that with the advent of the new era of new quality production factors, proprietary large language models are gradually converging, while open-source large language models continue to innovate, ushering in an era of "group model applications." For enterprise-level large language models targeting business applications, key issues during the implementation process include content compliance and credibility, model security and controllability, model data quality, deep integration of business, and project cost-benefit considerations.
From the perspectives of high-quality data, controllable generation, and security of intellectual creation, TRS-GPT LLM have integrated its independently developed Hybase Search Database, utilizing technologies such as RAG and knowledge graphs to combine large and small models with industry authoritative data sources, achieving compliant and controllable content generation. In terms of business integration, TRS-GPT LLM has developed an AI Agent application framework based on autonomously evolving task chains, combining knowledge and data-driven processes, enabling the autonomous construction of corresponding tool chains based on different input requirements, adapting to various application scenarios for Business users.
In addition, Lin Songtao also shared the case practices of TRS-GPT LLM in the media, finance, government affairs, public opinion, and public security industries, and further demonstrated the applications of intelligent knowledge question-answering, official document and manuscript writing, public opinion summarization, document review and proofreading, intelligent business assistants developed based on TRS-GPT LLM.
During the summit dialogue on "How large language models create future industrial value," Lin Songtao and the guests delved into the development trends of large language model technology, how to effectively utilize large language models, and maximize the value of large language models. He emphasized the need to build and utilize autonomous, controllable, and fine-tunable high-quality datasets for the precise application of large language models in specific fields. For large language models to be effectively implemented across various industries, continuous evolution of the model base and integration of business knowledge are required.
At the forum, CCID Consulting released the "Research Report on the Development of Large Language Models in China (2024)," selecting the top 20 leading enterprises in China's large language model industry in 2024 based on core technological capabilities, market competitiveness, product innovation capabilities, market development potential, and ecological openness capabilities. TRS was selected along with companies such as Baidu, Huawei, Inspur Information, and 360.
TRS was also listed among the top 100 artificial intelligence enterprises in China in 2024. As one of the earliest companies in China to develop artificial intelligence technology, TRS has over thirty years of experience in natural language processing research and leading technologies in the fields of NLP, knowledge graph, OCR, and structured image and video.
2024 marks the beginning of the era of large language model applications, with large language models blooming across various industries. TRS will continue to enhance its comprehensive research and development capabilities and key technological innovation capabilities to lead the digitization development with artificial intelligence and big data.
