Two researchers from China with no formal medical background were the winners of this year's Data Science Bowl, a competition held by consulting firm Booz Allen Hamilton and data science community Kaggle to facilitate the development of machine-learning algorithms through crowdsourcing.
This year's competition focused on using machine learning to improve the accuracy of CT lung cancer screening. Participants used anonymized high-resolution lung CT scans from a publicly available dataset provided by the U.S. National Cancer Institute to create algorithms designed to determine when lesions in the lungs are cancerous, according to Booz Allen. The winners from among nearly 18,000 submissions are as follows:
- First place: Liao Fangzhou and Zhe Li of Tsinghua University in China
- Second place: Julian de Wit and Daniel Hammack, software and machine-learning engineers from the Netherlands
- Third place: Team Aidence, members of which were from a Netherlands-based vendor that applies deep learning to medical image interpretation
The winners will split a $1 million prize, funded by the Laura and John Arnold Foundation. The winning teams presented their algorithms this week at the NVIDIA GPU Technology Conference in San Jose, CA.
