CDT Trip to Singapore 9-13 November 2019 by Juljan Krause and Allison Noble
When first- and second-year Web Science doctoral researchers Allison and Juljan disembarked the aircraft after a remarkably comfortable 13hr-flight on the A380, the biggest passenger plane in the world, they were immediately treated to another awe-inspiring engineering masterpiece. At 40 metres and surrounded by a four-storeyed terraced rain forest, the Rain Vortex at Changi Airport is the world’s largest indoor waterfall. Since its grand opening in April 2019, it has quickly become one of the most Instagrammed attractions in the world.
The waterfall certainly sets the scene for experiencing Singapore, this bustling melting pot that is home to 5.5m people – its diversity reflected by not one but four official languages: English, Mandarin, Malay and Tamil. It was difficult not to fall in love with Singapore right away. With its busy four-lane roads, buzzing alleys, Buddhist temples nestled amidst the high-rise buildings of the financial district, and above all, the amazing food. Average temperatures of 33°C proved a welcome break from the cold and rainy November nights back home in Southampton.
After a welcome rest day to shake off the jetlag, it was all hands-on deck at the Carlton Hotel. The NeXt++ Workshop kicked off on Monday 11 November with a talk from Dame Wendy Hall who reflected on the global issue of fake news and misinformation: true to the spirit of Web Science, Wendy highlighted the urgent need to consider the problem a complex socio-technical one that will require more than just the tweaking of some algorithms. In the following presentation, Mark Weal reflected on the role of social media data for interdisciplinary research. He argued for better consideration of data sources, bias and provenance in using such data (particularly Twitter feeds) and suggested some principles of best practice. Wendy and Mark’s talks were followed by a full day of varied and insightful presentations on the multifaceted ethical challenges that Artificial Intelligence and Machine Learning pose. Researchers from the National University of Singapore and Tsinghua University in China presented on topics such as explainability in neural machine translations and the neuropsychology of cheating.
On the second day of the NeXt++ Workshop (Tuesday 12th November), students from the University of Southampton, National University of Singapore and Tsinghua University were involved in a collaborative workshop which ran parallel to the remaining presentations. The aim of the workshop was to inspire an integrated effort in addressing one of the four following topics:
- Detection of Fake News
- Detection of Fake MM Contents
- Trust of Recommendation Systems
- Trust Calibration via Explanation
By having the opportunity to choose the project they worked on, Allison and Juljan found themselves to be working in the same group (project 3. Trust of Recommendation Systems). They were also joined by three students from Tsinghua University and the National University of Singapore. This newly assembled team took the time to introduce themselves and discuss their initial ideas surrounding the ideas of trust in recommender systems. After this initial brainstorming spell, the team agreed to include several key themes which were agreed as appropriate for the presentation; socio-technical issues, exploring previously attempted techniques, using explainability and robustness as a solution, and the proposal of a final framework encouraging trust. The team worked well together, putting their ideas into a shared Google document (meaning that everyone could see what the other was doing and could avoid overlapping of content). The team then presented their findings as a group to the rest of the workshop and found that they were met with interesting perspectives in response. The overall conclusion agreed by both the team and the rest of the workshop was that the matter of trust in recommendation systems was an issue which would be best resolved by a fusion of expertise which encompassed social and technical matters.
There is no doubt that the NeXt++ Workshop provided the visiting Web Scientists with valuable insight into the work being developed to solve various issues created in the continuous process that co-shapes technology and the social. It could also be suggested that this workshop provided the students from other establishments with some new perspectives on the importance of socio-technical study and how problems which seems purely mechanical can usually be aided with examination of the larger social issues surrounding development and use.