Deep Potential AI for Science Microscale Modeling System Wins Outstanding Achievement Award at the 2nd Zu Chongzhi Award

"Zu Chongzhi Award" – AI Frontier Innovation Award (referred to as "Zu Chongzhi Award") is a non-profit public welfare award initiated and hosted by the New Generation Artificial Intelligence Industry Technology Innovation Strategic Alliance (AITISA). This award aims to recognize outstanding achievements in fundamental research, frontier technologies, and innovative applications within the field of artificial intelligence.

In scientific research at the microscopic scale, traditional methods have long faced the challenge of balancing accuracy and efficiency. The algorithm system constructed through AI technology has overcome three key points in microscale modeling: structure, force field, and sampling. Taking the deep potential method as an example, it has successfully scaled molecular dynamics with first-principles accuracy and achieved the integration of machine learning, scientific computing, and high-performance computing, promoting efficient innovation in fields ranging from materials chemistry to biomedicine.

As the AI for Science field moves from single-point breakthroughs to the systematic platform construction phase, establishing an open-source and open ecosystem around microscale large models to provide new infrastructure for microscopic scientific research is gradually gaining consensus. The "Large Atomic Model" co-construction plan was launched at the end of 2023. Currently, more than 30 co-construction units have participated, and an advisory group consisting of several academicians has been formed. Through joint efforts within the industry, they are collaboratively addressing key challenges in atomic-scale research of complex systems, driving innovative breakthroughs in industrial microscale design in fields such as materials, energy, and biopharmaceuticals.

One of the goals of AI for Science in the next phase is to advance towards a new paradigm of the "Big Science Era." A key task moving forward is to build a complete chain that connects "reading (literature), computing (calculations), and doing (experiments)," achieving a "dry-wet closed loop" to systematically address pain points in scientific research scenarios, driving the evolution from a "toolkit" to an "intelligent agent." It is hoped that the construction of shared infrastructure within the industry, along with open-source approaches, will accelerate its implementation.

AI for Science is effectively helping to break down the boundaries between disciplines, the boundaries between theory and experiment, and the boundaries between scientific research and industry, enhancing exploration space and efficiency to address real problems in various industries. DP Technology will continue to focus on the "four beams and N pillars" construction approach of AI for Science to carry out technological innovation.AI for Science The co-construction and application of infrastructure, linking high-level, interdisciplinary talents, promoting more academic cooperation and exchanges, cultivating cross-disciplinary talents, creating more innovative opportunities, and driving the innovation of scientific research paradigms.