Lingnan University AI expert Prof Sam Kwong Tak-wu receives the 2026 IEEE Transactions on Evolutionary Computation Outstanding Paper Award.
The award, highlighting Prof Kwong’s international leadership in AI research, was established by the IEEE Computational Intelligence Society, whose journal is a globally recognised academic citation index database platform, ranked 11th among 204 journals in the Computer Science and Artificial Intelligence category of the Web of Science.
Published in 2023, the award-winning paper introduces a groundbreaking approach that integrates a learning component into traditional evolutionary computation methods, enabling AI to identify optimal solutions faster and more accurately in complex scenarios. The research team developed a novel, learning-aided, evolutionary optimization (LEO) framework, which allows artificial neural networks (ANNs) to learn from successful experiences during the evolutionary process. By observing which solutions outperform previous ones and recording this knowledge for future reference, evolutionary computation methods can move beyond random mutation to leverage past experience, significantly improving efficiency.
To evaluate the effectiveness of LEO, extensive experiments were conducted on international benchmark platforms, including single-objective and multi-/many-objective evolutionary optimization problems. Results demonstrate that LEO consistently outperforms traditional evolutionary computation methods, achieving higher efficiency and accuracy in complex problem-solving. Originally a highly forward-looking concept in 2023, LEO has increasingly been applied to areas requiring complex computations, such as traffic planning, smart manufacturing, drug discovery, and green energy, demonstrating broad practical potential.
Prof Kwong said “As global societies and industries face growing demands for rapid computation and precise prediction, traditional evolutionary optimization approaches often struggle with large datasets or multi-objective scenarios. Our innovative approach allows AI to learn from past experience, helping evolutionary optimization approaches avoid unnecessary steps and quickly find high-quality solutions. In today’s fast-evolving technological landscape, this methodology is becoming an essential tool in AI applications. This international recognition not only validates our team’s efforts, but also favours Hong Kong’s global leadership in AI research, with a long-term positive impact for society.”
The research paper is co-authored by Prof Zhan Zhi-hui and Dr Li Jian-yu from the College of Artificial Intelligence of Nankai University; Prof Sam Kwong Tak-wu, and Prof Zhang Jun from Hanyang University.
