When elephants make war, the grass suffers. But innovative folks will make hay with the trampled grass.
As European governments tighten the screws on Big Tech, three AI-obsessed students from SMU have found just such an opportunity to showcase their skills in the use of Artificial Intelligence to sniff out non-compliant products on a large online platform like AliExpress – potentially saving the platform a huge sum in fines.
Alibaba’s AliExpress, along with the likes of TikTok and X (formerly Twitter) have been scrutinised by the European Union under the new Digital Services Act (DSA) to weed out retailers with a history of non-compliance. As many small retailers must comply with varying restrictions before they can export their products to different countries, there is an urgent need for a cost-effective method to help these small business owners know, for example, if their thermal flask can be sold in China or another country.
Three second-year Computer Science students from the Singapore Management University (SMU) School of Computing and Information Systems have come up with such a scalable and automated solution, which won them the top prize at the Alibaba International Global E-Commerce Challenge held at Alibaba’s headquarters in Hangzhou on 4 September.
The second prize went to a team from Johns Hopkins University and the University of Maryland, College Park from the United States, while the third place went to a team fielded by Bandung Institute of Technology of Indonesia.
In the finals, there were 11 teams beside the SMU trio, including one from the National University of Singapore (NUS), one from Nanyang Technological University and one that had an NUS student combined with a Singapore Institute of Technology student. These 11 teams had been narrowed from 500 teams comprising over 1,100 participants from Europe, the Americas, East Asia and Southeast Asia.
How AliAdapt works
The solution by the SMU team is christened AliAdapt. It is an integrated system of three machine learning components that automate verifying whether a product listing can be exported to a certain country.
“For example, if a Singapore-based e-commerce business owner wishes to sell his thermal flask into China, our AI models can automatically classify his thermal flask product listing as exportable or not exportable into China,” said Joel Lim Jing Kai, (Year 2, BSc (CS)).
“If his product is exportable, the business owner receives a verification checkmark that allows him to expand into the Chinese market,” he elaborated. If the product is not exportable into the target country, AliAdapt continues to help the retailer out with obtaining permits to sell their product in the foreign country — usually a difficult process for e-commerce business owners.
Similarly, any product listing on AliExpress can be verified automatically by AliAdapt to be exported to any country. “Our solution also includes a dashboard that allows sellers on e-commerce platforms to easily get verified, and also enables them to track all of their product listings, as well as the countries they’re able to be sold in,” said Joel. “AliAdapt is designed to be scalable and cost-effective to adapt to the constantly changing regulatory environment.”
Creativity with computing
Joel’s teammates are Vince Tan Yueh Yang and Zane Chee Jun Yi (both Year 2, BSc(CS)). SCIS Dean Professor Pang Hwee Hwa said their victory is “a testament to the fact that our Computer Science students are not just adept with the computing concepts but also attuned to the business needs of companies.”
“Their creativity reinforces the value of the education and real-world experiences that we offer at SCIS and SMU in preparing our students to make meaningful impact wherever they go,” he added. “With the increased competition for tech jobs and internships, I am glad to see that our school culture encourages students to take the extra step to differentiate themselves.”
Even though the trio are just in Year 2, they are passionate about the use of Machine Learning (ML) to solve real-world problems.
They actively searched for AI and ML hackathons which could offer a springboard for postgraduate studies or ML engineering jobs in the future. “Alibaba's problem statement was the most interesting to, and the competition is a very prestigious one, so we decided to spend our time on it,” said Joel.
Have dreams will travel
Joel, Vince and Zane hold the SMU Global Impact Scholarship, which seeks to develop action leaders with a global perspective for the public, private and people sectors. The trio are involved in the SMU Business Intelligence and Analytics (BIA) club, and won US$10,000 in travel vouchers. As finalists, they will be given Alibaba International internship opportunities.
Joel is the head of SMUBIA’s 10th Data Associates Programme (DAP). All three participated in the DAP in Year 1, which gave them a strong grounding in ML to take part in the Alibaba competition. “We love to build working prototypes that have real-world impact. We saw an opportunity to make a global impact with this competition, so we signed up for it,” Joel shared.
As avid travellers, the team is thrilled with their prize money. “I take it as a sign to travel and expand my horizons,” said Zane, “the prospect of new adventures excites me.”
“Competing against such talented teams has pushed us to think creatively. The internship will be a great opportunity to work with some of the brightest minds in the industry,” said Vince. “We hope we can contribute to projects that have a real impact.”