12 February 2020
IoT Datathon 3.0
Terra AI and National University of Singapore (NUS) Faculty of Science got together again to co-organise IoT Datathon 3.0 during the school holidays in December 2019 which culminated in a finale event held on 17 January at Science Frontier Green. The finale event was sponsored by Schneider Electric, NVIDIA and SG Innovate.
The Datathon is a six-week programme with the objective to develop industry-ready talents capable of applying Artificial Intelligence (AI) and deep learning techniques for solving real-world problems. These skilled talents are in high demand by the industries around the world. The Datathon also serves as a platform for industry to explore and adopt solutions developed by the participants. The programme attracted more than 300 participants.
This year, the theme of IoT Datathon 3.0 was AI SAVE LIVES, where students had to develop AI-driven applications to predict one of the deadliest vector-borne diseases in the world – dengue.
Following its launch on 9 December 2019, the participants were grouped into more than 80 teams for the Training and Development Phase, to learn deep learning concepts and hands-on techniques applied by professionals for developing AI applications.
IoT Datathon 3.0 classroom training
CEO of Terra AI Mr Arnold Doray taught AI concepts to IoT Datathon 3.0 participants
Next, they proceeded to the Development Phase where they were given the datasets relevant to the challenge statements. The teams applied AI concepts learned during their training to develop AI models that can predict dengue outbreak in Singapore. The AI experts provided mentorship as necessary to nurture the students in their path to success.
IoT Datathon 3.0 students equipped and ready to develop dengue forecasting models
IoT Datathon 3.0 Challenge Phase climaxed at the Finale competition where the top eight teams battled it out. The teams presented their dengue forecasting applications to a panel of expert judges from Schneider Electric, NVIDIA, Terra AI, NUS, AI Singapore, and the National Environment Agency (NEA).
Teams presented their dengue forecasting solution to the judges and audiences
IoT Datathon 3.0 judges attentively evaluating the finalists’ solutions and presentations
In conjunction with the IoT Datathon Finale, the Data Science 2020 Industry Fair was organised with participation by top-notch hiring companies and organisations hungry for data science talents. IoT Datathon 3.0 participants became a talent pipeline for the companies offering AI and Data Science career opportunities.
The Champion team, Team MozzieNet developed a web-based machine learning application that can predict dengue outbreak, eight weeks in advance. Users can input desired parameters and instantaneously obtain the prediction results through a smartphone or laptop. The team members are Christopher GOH, Dennis TAN, Kaitlyn NG, SUN Yitao, TOH Wei Yang, a multidiciplinary team of students from Environmental Studies and Computer Science.
Team MozzieNet presenting their interactive web-based dengue application
Schneider Electric Zone President Tommy Leong and NUS Science Dean Prof Shen Zuowei presented the winning prize to IoT Datathon 3.0 Champion, Team MozzieNet, together with the other judges
The 1st Runner-up, Team Mopiko 2.0, developed a dengue forecasting application using a unique clustering approach exploiting various weather stations data geospatially spread throughout the country. The members are Clarence ONG, Benita NEO, CHENG Wenhao, Richard WONG, TOH Zi Yu, students from the Faculties of Science, Engineering and Computing.
The 1st Runner-up – Team Mopiko 2.0
The 2nd Runner-up, Team TFFL, developed a multi-model dengue predictor app with added interpretability features as well as strength and limitation analysis on each model. The team members are Ananya JOSHI, LI Xinyi, YU Peiran and ZHANG Wei, students from the Faculties of Science and Design & Environment.
The 2nd Runner-up – Team TFFL
Prof SHEN Zuowei, Dean of Faculty of Science, National University of Singapore, said, “The participants have put what they learned in theory to practice, picking up valuable skills that will give them an advantage in the marketplace. At the same time, I believe they have had fun in learning and pride in developing their applications. This is the best part about learning.”
Mr Tommy LEONG, Zone President of Schneider Electric and judge of IoT Datathon 3.0 said, “IoT Datathon is a catalyser for innovation and a breeding ground for the talent of tomorrow. Seeing finalists from different disciplines come together proved once again that diversity wins. This is in line with the talent philosophy we have at Schneider to drive innovation at every level.”
Mr Arnold DORAY, CEO of Terra AI, said, “The IoT Datathon was an eye-opener. Our students, some of them from non-computing backgrounds, successfully tackled a hard problem with huge social impact. A testimony to the capability of NUS students and I feel also validates our approach to teaching AI.”
Top 3 IoT Datathon winners with the judges and Prof Shen Zuowei
Team MozzieNet’s Kaitlyn Ng said, “IoT Datathon forces you to think deeper when solving a real-world problem. This is normally difficult to pick up on your own, especially for time-series problems which can be complicated. Being exposed to real-world datasets and problem statement gives you the experience you need for solving real-world challenges.”
Happy smiles from IoT Datathon 3.0 participants, partners and organisers