SCSE’s paper won the Innovative Application of AI Award from AAAI

SCSE’s paper titled “Efficient Training of Large-scale Industrial Fault Diagnostic Models through Federated Opportunistic Block Dropout” has won the Innovative Application of AI award from the Association for Advanced Artificial Intelligence (AAAI) on 9 Feb 2023. The paper was presented during the 35th Annual Conference on Innovative Applications of Artificial Intelligence (IAAI-23) held in Washington DC, USA from 7 to 14 Feb 2023 (https://aaai-23.aaai.org/iaai-23-call/). The IAAI Conference, through the Innovative Applications of AI Awards, honours case-study papers that describe deployed applications with measurable benefits that include some aspects of AI technology (https://aaai.org/about-aaai/aaai-awards/). 

Artificial intelligence (AI)-empowered industrial fault diagnostics is important in ensuring the safe operation of industrial applications. Since complex industrial systems often involve multiple industrial plants (possibly belonging to different companies or subsidiaries) with sensitive data collected and stored in a distributed manner, collaborative fault diagnostic model training often needs to leverage federated learning (FL). As the scale of the industrial fault diagnostic models are often large and communication channels in such systems are often not exclusively used for FL model training, existing deployed FL model training frameworks cannot train such models efficiently across multiple institutions.

The overall workflow of FedOBD.

To deal with such problems, the SCSE team led by Nanyang Assistant Professor Han Yu proposed the Federated Opportunistic Block Dropout (FedOBD) approach (Figure 1). By decomposing large-scale models into semantic blocks and enabling FL participants to opportunistically upload selected important blocks in a quantized manner, it significantly reduces the communication overhead while maintaining model performance. Since its deployment in collaboration with the ENN Group (https://www.enn.cn/) in February 2022, FedOBD has served two coal chemical plants across two cities in China to build industrial fault prediction models. It helped the company reduce the training communication overhead by over 70% compared to its previous AI Engine, while maintaining model performance at over 85% test F1 score. It is the first successfully deployed dropout-based FL approach. 

The team is thankful for the instrumental contributions by NTU SCSE research students and staff Mr Chen Yuanyuan (PhD, Year 2), Ms Chen Zichen (M.Eng, Class of 2020), Mr Zhao Yansong (PhD, Year 2), Dr Liu Zelei (former SCSE Research Fellow) and Dr Wu Pengcheng (NTU Research Scientist).

Mr Chen Yuanyuan (PhD, Year 2) receiving the award on stage

Team photo with the award during the AAAI-23 conference. From Left: Mr Chen Yuanyuan (PhD, Year 2), Nanyang Asst Prof Yu Han and Ms Chen Zichen (M.Eng, Class of 2020).

Prof Yu Han has also been conferred the title of AAAI Senior Member at the AAAI Award Ceremony on 11 Feb 2023. This year, a total of 15 honourees worldwide have been recognised: https://aaai.org/Awards/senior.php