From the source · Translated from Chinese
As the 'AI+' action deepens, accelerating the breaking down of barriers between basic research, technological breakthroughs, and industrial application has become a key path to cultivating new productive forces. On October 23, the 'Shanghai Gathering of Wisdom · Chip Research Symbiosis' Industry-Academia-Research Collaborative Innovation Exchange Conference and Intern National Science Exploration Challenge Award Ceremony, co-hosted by Shanghai AI Laboratory and Biren Technology, was successfully held. This event was a vivid practice of exploring collaborative innovation in the AI field. The Intern National Science Exploration Challenge, launched in July, attracted nearly 10,000 young talents from over 800 research institutions, universities, and enterprises worldwide. The competition collected over 400 Skills, all of which have been launched on the Intern·Duanyan Scientific Discovery Platform. Awards were presented to outstanding teams in six tracks. During the conference, Shanghai AI Laboratory announced multiple cooperation plans: jointly with East China Normal University, it launched the first phase of the 'National Wisdom Challenge'—the Digital Intelligence Solving Hackathon, and initiated the 'Intern·Duanyan' Science and Education Co-creation Plan; jointly with Biren Technology and Shanghai INESA, it launched a shared computing power pool plan, focusing on building an efficient and collaborative industry-academia-research innovation ecosystem. In the future, Shanghai AI Laboratory will further leverage its core functions as a 'connector, amplifier, and accelerator,' deepen the exploration of long-term mechanisms for industry-academia-research collaborative innovation, and promote the implementation of the 'AI+' action. Using the competition as a guide: opening up the industry-academia-research innovation transformation chain. In July 2026, Shanghai AI Laboratory, together with Biren Technology, launched the Intern National Science Exploration Challenge targeting real scientific research scenarios, setting up six tracks: life sciences, material sciences, earth sciences, quantum computing, operator models, and free exploration. The competition attracted nearly 10,000 participants, covering over 800 universities and institutions. Awards were presented to outstanding teams in the six tracks. Life Sciences Track: from AI design to cell experiment validation. The Shanghai University team, targeting the problem of candidate molecule discovery in liver lipid metabolism research, built an AI research agent, completed molecular structure design, target matching evaluation, and safety and stability screening, and finally proposed 10 candidate compounds. These compounds passed cell viability screening, with several significantly reducing triglyceride levels in cell experiments, some having novel structural features, forming a phased research result from computational design to experimental validation. Quantum Computing Track: generating verifiable scheduling schemes for multi-robot collaboration. The Shanghai Jiao Tong University team, targeting the problems of complex task allocation and resource conflicts in multi-robot collaboration, integrated large models with quantum optimization algorithms, converting natural language requirements into scheduling models containing task dependencies, resource, and time constraints. Relying on the UnitaryLab open-source algorithm library and capability library from Unitary Quantum and computing power support from Biren Technology, the team generated feasible solutions meeting constraints in test scenarios and verified them using classical solvers, providing a new solution path for complex collaborative tasks. Earth Sciences Track: exploring new clues for early identification of El Niño. The Xiong'an Meteorological Institute team, targeting the difficulty of early signal identification in El Niño prediction, used the EarthLink agent for human-machine collaborative research, expanding the analytical perspective from total ocean heat to the spatial structure of heat storage. The team built a 'ocean heat accessibility' mechanism framework, integrated multi-source ocean data and climate models for cross-validation, and identified signals leading the full heat release of the central Pacific by 3-5 years, providing new clues for exploring longer lead-time climate predictions. Material Sciences Track: making chemical reaction optimization more efficient and interpretable. Targeting the problems of high search costs and difficult-to-interpret optimization processes in chemical reaction conditions, multiple teams explored different solution paths.
shlab.org.cn
