Digitalization is bringing huge imagination to urban development, and at the same time, it is also promoting the reconstruction of the future digital city pattern. Standing at the critical stage of digital transformation, the Zhuhai Municipal Party Committee and Municipal Government have thoroughly implemented the strategic decision of the CPC Central Committee and the specific deployment of the "1310" of the Provincial Party Committee, and put forward the construction concept of "Smart City on the Cloud" in a forward-looking manner. Based on this, on June 22, at the Huawei Developer Conference 2024, Huawei Cloud Computing Technology Co., Ltd. and Zhuhai Unbounded Ark Intelligent Technology Co., Ltd. jointly released ArkModel-BioMed, a large model of Ascend cloud-native pharmaceutical R&D, and signed a comprehensive cooperation agreement to jointly accelerate the R&D and application of emerging technologies such as artificial intelligence and big data in Zhuhai, injecting surging impetus into the development of Zhuhai's new productivity and the strategic practice of "Smart City on the Cloud".
The first Ascend cloud-native pharmaceutical R&D model, activating the innovation engine of the biopharmaceutical industry
At the HUAWEI Developer Conference 2024 summit forum, Huawei Cloud Computing Technology Co., Ltd. and Zhuhai Ark Intelligent Technology Co., Ltd. jointly released the Ascend Cloud-native Ark basic model, ArkModel-BioMed, which is the first Ascend cloud-native pharmaceutical R&D model. The release of this achievement marks a huge impetus for the biomedical industry and injects new vitality into the innovation and development of the industry.
Compared with the general basic model, the ArkModel-BioMed model is built on the Ascend ecosystem, and HUAWEI CLOUD provides powerful computing power and basic cloud services. In terms of professionalism, the ArkModel-BioMed large model is injected with a large amount of professional text data in the training stage, including more than 250 million patents, 190 million papers, 860 million biological sequences, 250 million chemical structures, 40,000 targets and 80,000 drug data from 180 receiving offices around the world. At the same time, a large number of task inference logic data is innovatively added, covering a total of 17 subdivisions, including 5 million+ task inference logic trees, and the data is annotated in a semi-automatic way. In terms of security, the ArkModel-BioMed model adopts the industry's first "small and small brain" distributed model architecture, which can leave the data in the local server or device of the enterprise user for processing by marking the "privacy patch" when the user sends out a conversation, ensuring that the original text will not be leaked and can completely protect the privacy of user information.
During Huawei Developer Conference 2024, Zeng Xiaodong, Founder & Chairman of Zhuhai Ark Intelligent Technology Co., Ltd., delivered a speech titled "ArkModel Collaboration with ModelArts Studio to Promote the High-quality Development of the Biopharmaceutical Industry", introducing in detail the features of the ArkModel-BioMed model and introducing the world's first multi-expert agent drug R&D product jointly developed with industry partner Kemai Biotechnology (Suzhou). AgentStudio, and focused on the collaboration and innovation with ModelArts Studio. The AgentStudio platform is based on the ArkModel-BioMed model, and its innovation lies in the ability to combine multiple professional AI agents (Agents) with human expertise to form an efficient and collaborative virtual team. This hybrid orchestration of humans and AI agents not only reduces R&D costs, but also maintains the flexibility and innovation of the organization.
As the underlying technology support of the ArkModel-BioMed model, HUAWEI CLOUD's Ascend AI cloud service can support efficient innovation in 100 modes and 1,000 states, helping the ArkModel-BioMed model to be one step ahead of others in the development of the biopharmaceutical industry, and jointly promoting the innovation, upgrading, and digital transformation of the biopharmaceutical industry.
Working together, HUAWEI CLOUD and Ark Unbounded Ark have opened a new era of AI cooperation
During the Huawei Developer Conference 2024, Huawei Cloud Computing Technology Co., Ltd. also signed a comprehensive cooperation agreement with Zhuhai Ark Intelligent Technology Co., Ltd., which was attended by Li Zhuo, Partner and Commercial Director of Zhuhai Ark Intelligent Technology Co., Ltd., Zeng Xiaodong, Founder and Chairman, Cui Bin, General Manager of HUAWEI CLOUD Large Model Ecosystem, and You Peng, President of HUAWEI CLOUD Big Data and Artificial Intelligence Domain.
According to the cooperation agreement, the two parties will carry out in-depth cooperation in the co-construction of computing power bases and industry application cooperation. HUAWEI CLOUD will provide high-performance computing power based on the Ascend AI basic software and hardware platform, facilitate the training and iteration of the ArkModel-BioMed large model, and rapidly increase the scale of computing power through computing resource sharing, promoting the innovation and application of AI technologies. The two parties will also jointly explore the application of multi-agent products in different industries, develop more industry solutions, and share resources and technologies to promote the development and application of multi-agent products and drive industrial upgrading and digital transformation.
In the future, the continuous deepening of the cooperation between the two parties will effectively activate the huge potential of AI technology, including large models, in promoting the high-quality development of the industry, and will also give birth to more innovative solutions based on digitalization. At the same time, HUAWEI CLOUD will work with more partners and customers in Zhuhai to make every effort to develop new productivity in Zhuhai with the "Smart City on the Cloud" as the engine and help Zhuhai "develop new quality productivity according to local conditions".