Beijing Municipal Commission of Science and Technology, Zhongguancun Management Committee, Zhongguancun Science City Management Committee
Beijing General Artificial Intelligence Society, Zhongguancun Science City Artificial Intelligence Industry Committee, Beijing Tongzhi Artificial Intelligence Industry Innovation Center Co., Ltd., Dinghao Super Internet
Beta Intelligence (Beijing) Co., Ltd., Beijing Shengzhi Technology Co., Ltd., Beijing Bujiangjiu E-commerce Co., Ltd., Beijing Jiaxunfeihong Electric Co., Ltd
Beijing Institute of Technology, Beijing Normal University, Beijing University of Posts and Telecommunications, Beijing Information Science and Technology University
Guided by Beijing's industrial needs, utilizing competitive collaboration mechanisms, and aimed at solving practical enterprise problems
为畅通教育、科技、人才的良性循环，进一步推动高新技术产业的发展壮大，激发高校学子在人工智能领域创新的积极性、主动性，培育新质生产力，赋能高质量发展，海淀街道联合北京通用人工智能研究院，举办首届产学结合高校通用人工智能大赛。活动秉承“以北京产业需求为导向、以竞争协同机制为手段、以解决企业实际问题为目标”的思路，由企业根据自身人工智能实际需求出题，学生团队打擂揭榜，架设政企院校产学研深度融合的桥梁，以此更好服务首都发展。
To facilitate a healthy cycle of education, technology, and talent, further promote the development of high-tech industries, and inspire university students to actively innovate in the field of artificial intelligence, cultivate new quality productive forces, and empower high-quality development, the Haidian Subdistrict has partnered with the Beijing General Artificial Intelligence Research Institute to host the first University-Industry Integrated General Artificial Intelligence Competition. The event follows the principle of "oriented by Beijing's industrial needs, utilizing competitive collaboration mechanisms, and aiming to solve real enterprise problems." Companies propose challenges based on their actual AI needs, and student teams compete to tackle these challenges, building a deep bridge for industry-academia-research collaboration, thereby better serving the capital's development.
大赛分为产业需求组和科研需求组，参赛选手自行选择赛题，目标是根据出题单位实际需求，通过人工智能帮助出题党委解决产业实际场景问题或提出有建设性的解决方案。
The competition is divided into an Industry Needs Group and a Research Needs Group. Participants can choose their competition topics freely. The goal is to use artificial intelligence to help the proposing party solve real-world industry scenarios or propose constructive solutions based on the actual needs of the proposing organization.
大赛提供高性能算力环境，助力学生团队全心投入AI解决方案。
The competition provides a high-performance computing environment to help student teams fully commit to AI solutions.
The competition is open to all university students, regardless of age or nationality. Students enrolled in vocational colleges (secondary college), undergraduate, master's, doctoral, and postdoctoral programs are all eligible to register and participate by logging onto the official website or designated cooperative platforms.
第4~10名：获奖学生团体将获得优胜奖证书，
不排名次，鼓励高校学子“以解决企业实际问题为目标”，踊跃投身人工智能攻关第一线，培育新质生产力，赋能高质量发展。
4th to 10th Place: Award-winning student teams will receive Excellence Certificates, without ranking, encouraging university students to "aim at solving real-world enterprise problems" and actively engage in the forefront of artificial intelligence research and development, cultivating new quality productive forces, and empowering high-quality development.
除奖金、证书外，获奖选手还有机会获得出题单位实习、就业机会；“通计划”项目读博推荐机会；专业创业指导以及免费入驻专属孵化器。
In addition to prizes and certificates, award winners will have the opportunity to receive internships and employment opportunities from the question-setting organization; recommendation opportunities for doctoral studies under the "Tong Plan" project; professional entrepreneurial guidance, as well as free access to a dedicated incubator.
AI-assisted hearing optimization algorithm for complex dynamic sound fields, integrating non-linear acoustic computing and reinforcement learning
In complex dynamic acoustic environments such as streets, restaurants, and meeting rooms, hearing-impaired users often struggle to discern target speech due to ambient noise and multipath reflections. The participating teams need to design an AI-powered hearing aid optimization algorithm that combines nonlinear acoustic models to model the characteristics of sound wave propagation and utilizes reinforcement learning strategies to adaptively adjust noise reduction, echo cancellation, and gain parameters in real time, thereby enhancing the clarity of target speech and the user's listening experience.
In the process of evaluating AI sports education applications, a single AI smart camera is used to evaluate multiple people simultaneously in sports education projects, meeting the needs of sports teaching where multiple students can conduct physical tests simultaneously.
Gain model for online marketing based on demand forecasting for fresh retail with incomplete demand signal correction
In the fresh retail scenario, products pose a severe challenge to inventory management due to their perishable nature (such as short shelf life and high spoilage rate). To control spoilage costs, management generally adopts a conservative inventory strategy, which can easily lead to periodic stockouts. Sales data recorded during stockouts only reflect actual transaction volumes under limited supply conditions and cannot fully represent consumers' true demand. This truncation and distortion of demand signals make traditional time series forecasting models ineffective. This study aims to construct an efficient framework for demand forecasting in fresh retail that can adapt to incomplete demand signals.
More than 60% of China's railways are located in mountainous areas with complex geological conditions, where floods and debris flows occur frequently. According to statistics from the National Railway Administration, in 2024, floods and debris flows accounted for 72% of railway disruptions caused by natural disasters, resulting in direct economic losses exceeding 800 million yuan. Traditional manual inspections and point-based monitoring equipment suffer from delays in response and insufficient coverage, making it urgent to adopt intelligent solutions.
3D Scene Navigation Algorithms: Implementation of Point-to-Point Navigation Function in Various 3D Scenes
Navigation in 3D scenes is a very classic and worth exploring problem. We hope that participants can train an intelligent agent that can perform point navigation in various environments, utilizing RGB and depth maps to achieve point navigation functions.
2025年 2025
8.15~9.20
全国大赛启动 National Competition Launch
线上报名开启 Online Registration Opens
2025年 2025
9.11~9.24
初选筛查 Initial Screening
2025年 2025
11.5~11.16
企业级赛事较量 Corporate-level competition
十进三进行中 Top 10 to 3 in progress
2025年 2025
12月 December
线下决赛巅峰对决 Final offline showdown
北京·海淀线下见！ Beijing, Haidian offline meet!
Gathering university academic leaders and experienced corporate executives to provide professional evaluations of the competition outcomes from diverse perspectives
高校评委 University judges
企业评委 Corporate Judges
