The "most powerful brain" leading to the era of Scientific Intelligence 2.0, the Xinghe Qizhi Scientific Intelligence Open Platform is about to be launched.
If the characteristic of the "Science Intelligence 1.0 Era" was that a small number of scientists mastered considerable AI engineering capabilities to achieve phased results, then in the "Science Intelligence 2.0 Era", the scientific research field urgently needs more powerful and user - friendly infrastructure and collaboration platforms. This enables a large number of scientists to focus on scientific problems themselves and truly become the protagonists of exploration.
Against this background, the NovaInspire: Scientist - Centered AI Open Platform has emerged. Jointly developed by the Shanghai Institute for Science Intelligence, Infinite Light - Year, and Fudan University, this platform positions itself as an intelligent - agent - native full - chain science intelligence open platform centered around scientists. It provides global scientists with full - stack infrastructure covering high - value scientific data, open - source science intelligence models, efficient intelligent computing for scientific research, closed - loop wet and dry experiments, multi - agent reasoning and planning, and multi - disciplinary collaborative research, creating the "most powerful brain" in the Science Intelligence 2.0 Era.
The platform will be launched at the 2025 World Artificial Intelligence Conference and the High - level Conference on Global Governance of Artificial Intelligence (hereinafter referred to as WAIC 2025) to be held in Shanghai on July 26.
Six Capabilities: Covering the Entire Scientific Research Process, Autonomous Intelligent Agents Help Achieve the AI Power of Science
The NovaInspire Science Intelligence Open Platform builds a science intelligence ecosystem and drives the transformation of scientific research paradigms with six core capabilities:
Autonomous Scientific Exploration Engine: Through the collaboration of multiple intelligent agents such as data, models, and experiments, it realizes automated scientific reasoning, dynamic decision - making and planning, and closed - loop experiments, autonomously conducts scientific exploration, and improves the efficiency of scientific discovery across the entire chain. For example, AI Einstein is responsible for theoretical thinking, AI Fermi is responsible for experimental sensitivity, and AI Gauss is responsible for mathematical expression, realizing the collaboration of multiple intelligent agents in data, models, and experiments.
Universal Warehouse of Science Intelligence Models: It creates an open platform for scientific models that integrates multi - modalities and multi - disciplines. It has gathered more than 200 models from over 40 institutions in 12 fields, enabling the out - of - the - box use of the latest cutting - edge models and helping researchers build complex AI - driven scientific research chains at low cost and with low barriers.
Science Intelligent Computing Acceleration Platform: It builds a highly stable and efficient science intelligence computing power platform. With technologies such as intelligent fault - tolerance and continued training, dynamic model deployment, and GPU/CPU computing power fusion, it improves the efficiency of the research and application of cutting - edge scientific models: 99.7% of tasks are reduced from "hours" to "minutes", the number of deployed models with the same resources is doubled, and long - term training is achieved without interruption.
Closed - loop Wet and Dry Experiments: It realizes the integrated collaboration between AI simulation (dry experiments) and physical world experiments (wet experiments), achieving a closed - loop verification of model predictions in the real world. It has integrated the world's first cross - scale deep phenotypic measurement platform, intelligent laboratory autonomous intelligent agents, and automatic experimental dexterous hands, helping the full - process integration of scientific research from calculation to verification.
High - value and Trustworthy Scientific Data: It constructs a scientific data ecosystem based on blockchain. The platform has gathered more than 40,000 scientific data subsets, with a total data volume of 12PB, a daily collection volume of 150TB, and a daily processing volume of 50TB. It fully supports the trustworthy collection, intelligent quality control, and collaborative sharing of data, providing solid data support for science intelligence.
Multi - disciplinary Cross - collaborative Research Community: Facing major scientific problems, it establishes a matching mechanism and an incentive system to aggregate diverse forces such as scientists, AI developers, and engineers. Scientists lead research planning and strategy design, data engineers are responsible for efficient data processing and computing power optimization, algorithm engineers build scientific models and analyze data laws, and experimental engineers conduct verification experiments, provide real - world data and feedback, and complete the verification closed - loop.
Three Features: Scientist - centered, User - friendly, and Co - built and Shared
When the technical foundation and collaboration network are ready, the spotlight of NovaInspire will focus on the real core - scientists. The powerful core of the NovaInspire platform stems from its six deeply integrated core capabilities, which are condensed into three distinct features:
Scientist - centered: Instead of making scientists "switch to AI", it makes AI serve scientists. Scientists focus on "asking the right questions", and AI helps "find the answers".
User - friendly: No complex deployment is required, and it can be used without learning. The platform is ready - to - use, with intuitive operations. Research ideas can be implemented with one click, and everyone can use AI for scientific research.
Co - built and Shared: It returns to the essence of science, with open - source and collaborative cooperation. It breaks down the barriers between scientific research institutions and disciplines and co - builds an open and cooperative global science intelligence innovation network.
The three features are implemented through the six capabilities, enabling scientists to naturally control powerful intelligent agents, truly realizing "ScienceAI" - the AI power of science, effectively releasing the productivity of artificial intelligence, accelerating scientific discovery, and creating the "most powerful brain" in the Science Intelligence 2.0 Era. On the stage of NovaInspire, the protagonist has never changed. Scientist is the star.
Being scientist - centered is not only a technical philosophy but also the core symbol of the Science Intelligence 2.0 Era. In the Science Intelligence 1.0 Era, AI was like a pruning shear, which required researchers to use it unilaterally to pick relatively mature "scientific research fruits". In the Science Intelligence 2.0 Era, AI has become a full - process "scientific research partner", accompanying and co - building throughout the process from inspiration to experimental verification, helping scientists directly address those profound "scientific questions" in the galaxy.
As we move towards a new era of science intelligence, the NovaInspire Science Intelligence Open Platform sincerely invites global scientists, AI engineers, and developers to pay attention and participate, and jointly embark on the golden age of science intelligence. The detailed functions of the platform and the first batch of settled projects will be announced at the conference on July 26. Please stay tuned for more information on the official WeChat account of the Shanghai Institute for Science Intelligence.