中文BioGeometry, a digital biology pioneer company that specializes in applying generative AI to protein design, recently announced a strategic partnership with Zhipu AI dedicated to constructing a large multimodal model（LMM）that aligns human natural language with the language of life. This model is expected to enhance the practical applications of generative AI platforms in the realms of life sciences and medical research, lower the barriers of entry for users, and efficiently process biomedical information. It will inspire users to discover new targets, molecules, and questions, providing novel ideas and tools for big molecule drug discovery with AI, and advancing biotechnology and pharmaceutical technologies.
Through this collaboration, BioGeometry and Zhipu AI will integrate their respective technological and business strengths to jointly develop a LMM that aligns knowledge in natural language with the language of life. This will allow for the intelligent translation and bridging of human language with biological information, focusing on mining answers to questions in a more direct and efficient manner for information retrieval. Users will be able to conduct iterative inquiries in natural human language, gaining intelligent assistance in research processes such as literature searches and molecular discovery, thereby solving complex challenges and constructing vertical large models in the field of life sciences, becoming powerful tools in the exploration of the mysteries of life sciences.
Dr. Jian Tang, founder and CEO of BioGeometry, stated: “We are delighted to establish this strategic cooperation with Zhipu AI. By combining our rich experience in AI-driven drug development with Zhipu AI’s impressive achievements in the domain of general large models, we are confident that our joint research will achieve groundbreaking progress in life sciences and medicine research.”
Zhipu AI also expressed its anticipation for the cooperation: “BioGeometry has accumulated a wealth of experience in the generation and understanding of molecular 3D structures, while we possess unique technical capabilities in natural language and multimodal data integration. Through deep cooperation, we expect to fully leverage our strengths, create vertical large models that can address complex issues in the biomedical field, and bring new momentum to human health initiatives.”
Currently, generative AI technology has made significant advancements in tasks such as natural language understanding, as well as image and video comprehension, paving the way for the realization of general-purpose AI. The rapid development of biotechnology has produced a large volume of heterogeneous and multimodal data, including molecular-level data such as gene sequences and protein structures, as well as biomedical literature. Meanwhile, the diversity and complexity of this biological data pose challenges for biopharmaceuticals. Applying AI technology to the biomedical field represents the strong combination of both technologies. It provides not only innovative solutions for current complex problems in biopharmaceuticals, but also the potential to address deeper biological questions in the future, contributing more innovative results to human health.
As a leading AI biotech enterprise, BioGeometry has extensive experience in AI drug development, and has collaborated with companies such as NVIDIA, Intel, IBM to successfully develop and open-source the machine learning systems specially designed for drug research — TorchDrug/TorchProtein available at https://torchprotein.ai/. It recently released the world’s first generative AI platform for antibody discovery, GeoBiologics, accessible at https://geobiologics-cn.biogeom.com. Zhipu AI, a leader in the field of general large models, has launched the 100 billion parameter super-scale pretrained model GLM-130B along with related dialog models, showcasing its strength in high-performance hundred-billion-scale large model technology.
About BioGeometry
BioGeometry an AI-driven biotech company dedicated to understanding the language of life through generative AI (AIGC), reconstructing the protein drug discovery and design process, and creating programmable proteins. It solves issues related to the traditionally high costs, long cycles, low success rates, and difficult drug targets in large molecule pharmaceuticals, providing innovative drugs to save patients’ lives. Founded in 2022 by Dr. Jian Tang, a renowned scientist in AI4Drug and deep learning on graphs, with Turing Award laureate Yoshua Bengio serving as the chief scientific advisor, the company has built two foundational platforms: an AI-driven large molecule drug design platform and high-throughput wet-lab validation, creating a closed-loop system that empowers partners to achieve precise and efficient large molecule drug design and optimization.
Business Cooperation: bd@biogeom.com
Official Website: www.biogeom.com
About Zhipu AI
Zhipu AI is dedicated to building the next-generation cognitive intelligent large models, with a focus on Chinese innovation in large models. The company has co-developed a bilingual 100 billion-scale super-large pre-trained model GLM-130B, launched a 100 billion-scale dialogue model ChatGLM and an open-source single-card model ChatGLM-6B, and has created a large model product matrix, including generative AI assistant ZhiPu QingYan, high-efficiency code model CodeGeeX, high-precision text-to-image generation model CogView, and multimodal dialogue language model VisualGLM-6B, etc. The company practices the market concept of “Model as a Service” (MaaS) and has launched a large model MaaS open platform. Based on the leading 100 billion-scale multilingual and multimodal pre-trained models, it realizes a new paradigm of AI development – high-efficiency and general-purpose “model as a service”.
Business Cooperation: service@zhipuai.cn
Official Website: zhipuai.cn