18 November 2020 Department of Computer Science , Faculty , Student , Alum

18 November 2020 – Associate Professor Kan Min-Yen and Computer Engineering alumnus Nguyen Van Hoang won the Best Paper Full Research Paper Award at the 29th ACM International Conference on Information and Knowledge Management (CIKM2020), held online from 19 to 23 October.

A/P Kan and Nguyen, together with co-authors Dr Preslav Nakov from the Qatar Computing Research Institute, and Associate Professor Kazunari Sugiyama from Kyoto University, were awarded for their work ‘The FActual News Graph’, or FANG, a new framework that uses Graph Neural Networks to detect misinformation on social media.

FANG works by modelling the spread of news on social media as a large network, incorporating information about the news and various news sources, such as mainstream media sites and ordinary online users.

Nguyen and A/P Kan said that their work probably stood out because it fit the theme of CIKM 2020, which was ‘data and knowledge for the next generation: sustainability, transparency and fairness’.

“Unlike previous work which focuses solely on improving the accuracy of fake news detection, our approach is both accurate and transparent as it meaningfully represents news coming from both social media users and traditional media sources,” said Nguyen.

A/P Kan added: “By taking into account the entire ecosystem of the news article's natural language, historical trends regarding the publishing source, as well as both proponents and detractors that comment and spread the news, our work helps to transparently assess the news and its sources.”

In order to collect the data required for the framework, they had to crawl through hundreds and thousands of social media profiles, user interactions, and the contents of media sources.

“It was challenging to conduct these large-scale experiments, as we required specialised computing hardware to process graphs containing a few hundred thousand data nodes,” said Nguyen.

Thankfully, they were able to manage the challenges with the modeling and resources needed to execute the experiments with help from the school and Amazon Web Services, said A/P Kan.

Their hard work and effort has paid off – especially for Nguyen, who previously also won the NUS Outstanding Undergraduate Researcher Prize for this work.

“I am very happy and thankful for the recognition from NUS and the data mining research community for our work. I hope this recognition can inspire other undergraduates that regardless of being an undergraduate or graduate, we can always publish impactful research and will be recognised for our competency,” said Nguyen.