Global AI Challenge for Building E&M Facilities presents Gold Award to Lingnan’s AI Model predicting cooling demand in commercial buildings

Hong Kong’s commercial buildings are often perceived as being excessively cold, which contributes to an unnecessary waste of energy and carbon emissions. A doctoral research team from Lingnan University’s School of Data Science (SDS) has developed an artificial intelligence (AI) model to predict cooling loads precisely, thereby significantly improving energy efficiency. The project recently won the Gold Award at the Global AI Challenge for Building E&M Facilities 2025, and the technology has already been tested successfully in a commercial building in Hong Kong and in the Das Intellitech Building in Shenzhen, both trials showing a strong predictive performance.

The award-winning project “Dynamically engineered multi-modal feature learning for predictions of office building cooling loads” was supervised by Prof S. Joe Qin, President and Wai Kee Kau Chair Professor of Data Science at Lingnan University, and Prof Mo Yanfang, Assistant Professor of the Division of Industrial Data Science of SDS, and team members included Dr Liu Yiren, Postdoctoral Fellow of the Division of Industrial Data Science of SDS, and PhD students Mr Li Jicheng and Mr Zhu Zhongxi from SDS. For the competition, the system accurately predicted the hourly cooling load requirements of four buildings in the month ahead. It also provided optimised control strategies for the HVAC (Heating, Ventilation, and Air Conditioning) systems, maintaining comfortable indoor temperatures, and reduced energy consumption and carbon emissions. The project received the Gold Award of the “Build Your AI Model” academic group, and a prize of US$5,000.

President S. Joe Qin said “This award highlights Lingnan University’s unwavering commitment to advancing the United Nations Sustainable Development Goals through innovation and technology. The University has consistently integrated the concept of sustainability into its teaching and research, while also showcasing its interdisciplinary strengths in data science and sustainable development. We look forward to deepening our collaboration with industry to leverage AI technologies for green building, contributing solutions to the global carbon neutrality goals.”

Dr Liu Yiren added “Most commercial buildings in Hong Kong currently rely on experience-based or fixed schedule adjustments for cooling supply, often leading to overcooling and energy waste. Our AI system can forecast hourly cooling load demands up to one month in advance, enabling significant energy savings without compromising occupant comfort. The model also possesses broad applicability, accurately predicting cooling needs across different building types, helping commercial properties reduce energy expenses substantially. We aim to expand the system’s adoption in more shopping malls and office buildings, promote the popularisation of smart building technologies, and contribute to Hong Kong’s carbon neutrality goals by 2050.”

The system has been evaluated in a commercial building in Hong Kong and Das Intellitech Building in Shenzhen to collect data, further validating its outstanding capabilities. Building on this technology, the team has developed an “intelligent HVAC management platform”, which integrates real-time weather data to predict cooling load demands under various conditions, and provides a visualised interface to assist building managers in formulating energy-saving strategies.

The team is now working on integrating the platform with large language models (LLMs) and AI agent technologies, and on enabling the system to proactively suggest practical energy-saving measures directly to building managers. This advancement aims to lower technical barriers and accelerate the effective implementation of smart building solutions.

Organised jointly by the Electrical and Mechanical Services Department (EMSD) and the Guangdong Provincial Association for Science and Technology, the Global AI Challenge for Building E&M Facilities was held from June to August 2025. Its two competitions, “Build Your AI Model” and “Innovative Proposal for Construction and Engineering”, attracted 200 teams from 26 countries and regions worldwide.