2024 Inclusion Bund Conference AI Innovation Competition and Deepfake Challenge Results

The 2024 Inclusion Bund Conference officially opened on Thursday in Shanghai, and on September 3 the organizing committee released the final results of the newly added "Technology and Intelligent Innovation Competition," which comprises the AFAC2024 Financial Intelligence Innovation Contest and the Global Deepfake Attack‑Defense Challenge.

Guided by the Shanghai Municipal Science and Technology Commission, the competition attracted more than 7,000 teams from over 20 countries and regions, including top universities such as Peking University, Fudan University, Stanford University, MIT, and NUS, as well as leading companies like Microsoft, Google, Meituan, and Xiaohongshu, and research institutes such as the Chinese Academy of Sciences Automation Institute, Microsoft Asia Research, and Singapore Institute of Technology.

In the AFAC contest, six outstanding teams earned first‑place awards in the Challenge and Enterprise categories and second‑place awards in the Startup category (the Startup first‑place slot remained vacant). The Deepfake contest saw a team from the University of Macau and individual contestant Tang Yongwei win the image and audio‑video tracks respectively.

Hong Kong City University post‑doctoral researcher Wu Haiwei led a Hong Kong‑Macau joint team to win the Deepfake image track, achieving a multi‑scenario deep‑fake detection rate of 97.038%.

Individual contestant Tang Yongwei also secured the Deepfake video track championship, earning a second‑place award in the AFAC Financial Tools Learning track, thus becoming a "double champion" by leveraging AI algorithms to detect forged audio and video.

AFAC participant and Shopee algorithm engineer Ma Qianli emphasized that the competition helps differentiate candidates in job markets by demonstrating innovative and practical abilities.

The event’s philosophy of "promoting innovation through competition" provides a platform for diverse talent to showcase skills and apply technology to real‑world problems, fostering interdisciplinary talent development.

The competition offered a dataset exceeding one million entries, covering over 100 fake‑image generation methods and more than 100 audio‑video attack combinations, bridging theory and industry applications across sectors such as healthcare, fintech, and smart manufacturing.

AFAC contest categories include Challenge (for developers using real fintech data), Startup (for entrepreneurial teams designing commercial solutions), and Enterprise (for SMEs proposing innovative fintech projects).

The Deepfake Attack‑Defense Challenge, chaired by Chinese Academy of Engineering academician Wang Yaonan, featured problem statements from Ant Group's ZOLOZ and Tianji Lab, making it one of the most authoritative events in computer vision.

Ant Group CEO Zhao Wenbiao highlighted the global significance of deepfake threats and the competition’s role in advancing industry‑wide security defenses.

According to Singapore Institute of Technology’s Deputy Director Zhou Tianyi, the competition provides researchers with realistic industrial scenarios, promoting academia‑industry collaboration and responsible talent cultivation.