On December 5, 2025, the event of Visual Computing and Simulation Specialty Committee of CSF visited university and The Eighth Scholar Forum of the School of Computer and Electronic Information, Nanjing Normal University, sponsored by the Specialty Committee and hosted by the School of Computer and Electronic Information of Nanjing Normal University, was successfully held in the Fangfei Building conference room of Nanjing Normal University. The forum, themed "Multimodal Intelligence and Large Model Applications", focused on discussions about cutting-edge technologies such as multimodal intelligence, embodied perception, and large models.

This forum invited Professor Linwang Yuan, Vice President of Nanjing Normal University, Professor Weishi Zheng, Deputy Director of the Visual Computing and Simulation Committee of the Chinese Society for Simulation, and Professor Jianjun Qian, Executive Dean of the School of Computer and Electronic Information at Nanjing Normal University, to deliver speeches. President Yuan introduced the development layout and thinking of Nanjing Normal University in the field of artificial intelligence. Director Zheng introduced the significance and planning of the establishment of the Visual Computing and Simulation Special Committee. Dean Qian expressed sincere gratitude to the attending leaders, guests, teachers, and students, and reported on the achievements of the college in disciplinary construction and other aspects. This forum is hosted by Wanqi Yang, Vice Dean of the School of Computer and Electronic Information/School of Artificial Intelligence at Nanjing Normal University.

This forum invited three authoritative experts and scholars in the field of computer and artificial intelligence, namely: Professor Weishi Zheng, Deputy Director of the Special Committee, Changjiang Scholar, and Vice Dean of the School of Computer Science at Sun Yat sen University; Professor Jun Yu, National Outstanding Youth and Dean of the School of Intelligent Science and Engineering at Harbin Institute of Technology (Shenzhen); and Professor Chen Tian, National Outstanding Youth and Dean of the School of Computer Science at Nanjing University..

Professor Weishi Zheng gave a lecture on "Embodied Perception and Learning for Mechanical Agile Hand Operations", and combined it with the trends in the field of robotics to point out the core value of interaction in mechanical dexterous hands. At the same time, he pointed out two major challenges: firstly, data collection is difficult in traditional remote operation mode, and there are problems such as mismatch between human-machine joints and low precision in complex operations; Secondly, model learning is constrained by the high degree of freedom of dexterous hands, the tendency for deviation between simulation and actual structural mapping, and insufficient sample size, which affects training. Subsequently, he presented the team's research findings.

Professor Chen Tian gave a presentation on "Intelligent Fault Location Based on Large Models in Production Networks". He first pointed out the pain points in the industry: Alibaba Cloud's global network scale has reached 100000 levels, and traditional automation tools can only solve 90% of faults, with the rest requiring manual processing. There are problems such as complex data, expert high-pressure misjudgment, and long-time consumption. Subsequently, he introduced the core solution, a hierarchical multi pipeline system: the large model first extracts log information, then analyzes and determines the probability of failure through single device analysis, and finally combines spatiotemporal analysis to sort suspected devices, focusing on reducing the burden on experts.

Professor Jun Yu presented a report titled "Multimodal Perception and Autonomous Decision making for Low Altitude Tailored Intelligent Agents". Firstly, he pointed out the application value of low altitude embodied intelligent agents in agricultural inspection, emergency rescue and other scenarios. At the same time, he pointed out the challenges faced: difficult fusion of multimodal perception data, insufficient robustness of autonomous decision-making in complex environments, and further increased technical challenges due to airspace limitations and dynamic obstacles in low altitude scenarios. Subsequently, he elaborated on the team's research findings: constructing a multimodal perception fusion framework, integrating visual, LiDAR, and other data to achieve precise environmental modeling. In addition, Professor Yu also introduced the application of technology in the low altitude inspection prototype system, verifying the effectiveness of perception and decision-making collaboration. Finally, he looked ahead to the prospects of the field。

This forum set up the academic exchange session for young scholars. Ying Li, Bo Zhang and Xudong Tian from the school of computer and electronic information of Nanjing Normal University reported their research results around "fine-grained image retrieval", "knowledge map enhanced intelligent education" and "the simplest and full visual representation learning". Associate Professor Ying Li reported on the breakthroughs made in art and design image retrieval and large-scale cross modal retrieval. Aiming at the pain points of the application of traditional large model education, Dr. Zhang Bo proposed a closed-loop cognitive intervention from establishing standards, accurate diagnosis to enlightenment guidance, and accelerated the implementation of intelligent evaluation of science experiments in middle schools. Dr. tianxudong focused on the simplest and full learning, and proposed an anti-intuitive scheme for the pain point of the information bottleneck theory, expanding it to multi view complex scenarios and achieving performance improvement。