With the rapid advancement of artificial intelligence, AI has achieved human-comparable performance in cognition and perception. However, its long-term vision extends beyond replicating human knowledge and skills—it aims to assist in exploring the unknown, accelerating scientific discovery, and transforming the world. AI for Science (AI4S) has emerged in this context as the “Fifth Paradigm of Scientific Discovery”, integrating AI deeply into various aspects of scientific research. AI4S is set to accelerate humanity’s understanding of nature, addressing critical challenges in medicine, energy, and materials science.

The 2024 Nobel Prizes in Physics and Chemistry were awarded to scientists who made groundbreaking advancements in AI, underscoring AI’s pivotal role in scientific discovery and marking a new phase for AI4S development.

• Advancing AI-Science Integration: Promote AI applications in scientific research to enhance efficiency, quality, and automation in scientific discovery.
• Building Foundational Scientific Models: Encourage participants to develop models capable of processing high-dimensional, multimodal data to learn and understand scientific concepts, entities, and principles, significantly boosting research efficiency.
• Enabling a Closed-Loop Scientific Workflow: Support the full cycle of scientific discovery, from theoretical formulation and computational simulation to experimental validation, driving automation across the entire research chain.
• Democratizing Scientific Exploration: Lower the barriers to scientific discovery, enabling more people to engage in research through AI tools, fostering a widespread enthusiasm for scientific exploration.

To achieve these goals, DI-IDEA, relying on Peking University, the secretariat unit of the DI-IDEA alliance, and collaborating with DP Technology, AI for Science Insitute, Beijing, Wuhan University, and LeadArt Biotechnologies, has introduced the AI for Science track as part of the 2nd Global Digital Intelligence Education Innovation Competition.

The AI for Science track features three key challenges:

Intelligent Experimentation

This topic focuses on the automation and intelligent control of organic chemical synthesis. Participants will use laboratory control software, provided components, and experimental protocols to design and implement automated synthesis workflows.

Geoinformatics

This topic invites participants to explore innovative applications of aerospace information combined with AI and big data. Teams will define their own problems and propose creative, feasible solutions with societal impact. The focus is on advancing AI-powered aerospace data processing and analysis in areas such as agriculture, forestry, transportation, and urban governance.

High-Performance Computing

This topic focuses on the high-performance optimization of large AI for Science models, addressing challenges in adapting and optimizing them for various hardware platforms. Participants are required to deploy workflows on domestic hardware, optimizing algorithms or adaptation strategies to enhance performance while maintaining a certain level of effectiveness.

Through the AI for Science track, we aim to bring together global expertise to drive AI-driven innovation in science, accelerate the pace of discovery, and contribute to the advancement of human society.

1.Participants must be full-time undergraduate, master’s, or doctoral students at institutions of higher education, or individuals who graduated from or left such institutions within the last three years (after January 1, 2022). The status of being a student is determined as of the official announcement date of the competition.

2.Each participating team may have no more than ten members, with no more than five members from the same institution of higher education. Each individual may join only one project in the AI4S Track.

3.Each participating team can choose whether to appoint instructors, with a maximum of three allowed.

4.Detailed personal identification information, team details, and instructors information will be collected in the semifinal stage for eligibility verification.

5.During the competition, if any submitted materials are found to be false, or if any participants are caught plagiarizing or violating others’ intellectual property, the participants involved will be disqualified.

6.Personnel involved in topic compilation and those with access to data from the competition organizers and technical support teams, as well as their close relatives, are prohibited from participating.

Registration

• Registration Start: April 3, 2025
• Registration Deadline: May 31, 2025

Preliminary Round

• Preliminary Round Start: April 3, 2025
• Preliminary Round Deadline: June 15, 2025

Semi-final Round

• Semi-final Round Start: June 30, 2025
• Semi-final Round Deadline: August 31, 2025

Final Round

• Scheduled between September and October 2025 (exact date to be announced)

Award Ceremony

• Early November 2025 in Beijing, China (specific date and format will be announced separately)

All the times mentioned follow Beijing Time (GMT+8).

The competition committee reserves the right to adjust the rules and schedule if necessary.