With trusted knowledge services to promote the implementation of AGI, "Caizhi Technology" has completed a tens of millions of yuan angel round financing led by Zhipu. | This is the exclusive first report by 36Kr.
Written by Wang Fangyu
Edited by Su Jianxun
36Kr has learned that Beijing Caizhi Technology Co., Ltd. (hereinafter referred to as "Caizhi Technology") has recently completed an angel round of financing of tens of millions of yuan, led by Z Fund (Zhipu Ecological Fund) and followed by Shengjing Jiacheng.
Caizhi Technology is an AI veteran player that has long focused on the research of trusted AGI (General Artificial Intelligence), and is committed to the application and implementation of interpretable and traceable large-scale models. In 2021, it jointly undertook the major project of the Ministry of Science and Technology's "New Generation Artificial Intelligence in 2030" with the Beijing Institute of Artificial Intelligence, and has rich accumulation in the research of large-scale model application technology.
In terms of the team, Caizhi Technology is also a typical "Tsinghua Department". Among them, the founder and CEO Dr. Xu Jianjun and the chief researcher Li Qi are all from the Knowledge Engineering Laboratory (KEG) of Tsinghua University, and have the same origin as the Zhipu Department.
The Knowledge Engineering Laboratory (KEG) of Tsinghua University put forward the research direction of "Knowledge Processing in the Network Computing Mode" 25 years ago; the "Shenzhi Intelligence" special knowledge model created by Caizhi Technology, which took four years to aggregate a large amount of knowledge data of normative documents in various industries, integrates the large knowledge engineering with the large language model, and becomes a ready-to-use, large-model-native trusted knowledge service "super" tool, helping various industries to achieve the safe and reliable implementation of large-scale model applications.
At present, the AI large-scale model has entered the stage of commercial application, and is empowering the transformation and upgrading of thousands of industries. However, the lack of credibility has become a major constraint for the landing of the current large-scale model business scenarios. At present, the conventional large-scale models and agent technology paradigms in the market still cannot break through the shackles of the large-scale model hallucination. If the AGI application and the agent cannot guarantee the accuracy, let alone generate actual business value, it is difficult to penetrate into complex and low-tolerance industry scenarios.
"The various difficulties faced by the application of large-scale models in the industry stem from a basic first principle of knowledge engineering. Common knowledge such as human common sense, science, language, and social encyclopedias is widely distributed in the literature data of all mankind; by training it with the Transformer self-supervised algorithm combined with huge computing power, the large-scale model can master the cognitive ability of common knowledge. But on the other hand, the knowledge of various industries is often knowledge that is authoritatively released, artificially set, and not naturally formed. In principle, it cannot be completed by a large-scale model alone for trusted cognition." The company's founder and CEO Xu Jianjun told 36Kr.
Trusted AGI = Large Language Model + Large Knowledge Engineering Image Source: Authorized by the Enterprise
Based on this technical insight, the "Shenzhi Intelligence" special model of Caizhi Technology embeds the full amount of ontological cognitive graphs and the reconstructed and woven massive knowledge data of various industries. Through supervised training and integration with large-scale models, it strictly guarantees 100% business traceability, dynamic accuracy and process interpretability of the AI system output.
Caizhi Technology uses the ready-to-use trusted intelligent knowledge service to assist in the standardized and rapid delivery of large-scale model applications, enabling the intelligent agent to go online quickly and accurately empower the scenarios.
Up to now, Caizhi Technology's products have been recognized by many large government and enterprise customers. It has directly signed contracts and created many benchmark cases for important institutions such as the General Office of the State Council, the PLA News Center, and Beijing Enterprises Group. It has achieved profitability last year. Xu Jianjun said that Caizhi Technology hopes to work with a wide range of partners to integrate the trusted AI knowledge service capability into every intelligent application, so as to truly empower thousands of industries and facilitate the public life.
In the next step, Xu Jianjun told 36Kr that the cognitive intelligence revolution driven by the large-scale model will trigger a market explosion in the field of trusted knowledge services, and 100% trustworthiness will become an indispensable key feature. Caizhi will take this opportunity to promote the improvement of the entire industrial ecology and contribute to the development of new-quality productive forces.
Investor's Perspective:
Z Fund (Zhipu Ecological Fund): In the new era that attaches importance to new-quality productive forces, serving the production and operation of B-end customers, reducing costs and increasing efficiency is a crucial part of the practical application of large-scale model technology, which involves multiple aspects such as data sorting, knowledge engineering, efficiency optimization, and business logic precipitation. We are optimistic about the product accumulation, technical capabilities and industry experience of Caizhi Technology. Caizhi builds truly usable, trusted and reliable large-scale model government and enterprise solutions and standardized products based on the Zhipu GLM series base large-scale model. We believe that as one of the first batch of investment enterprises of Z Fund, Caizhi Technology's products can effectively promote the overall process of large-scale model commercialization and promote the prosperity of the overall innovative application ecosystem.