MORE FROM KrASIA

Alibaba Cloud open-sources its machine learning algorithm platform Alink

The platform has been deployed in powering various businesses within the Alibaba ecosystem.

Photo: Shutterstock.com

Chinese e-commerce giant Alibaba’s subsidiary computing arm Alibaba Cloud announced Thursday that it has open-sourced the core codes of Alink, its self-developed algorithm platform, according to a press release of Alibaba.

Alink, which is built on Apache Flink, an open-source framework for managing big data at major scale, has already been deployed in powering various businesses within the Alibaba ecosystem. For example, it helped Alibaba to increase the click-through rate of product recommendations by 4% on the e-commerce platform Tmall during this year’s Singles’ Day shopping festival, said Alibaba.

“As a platform that consists of various algorithms combining learning in various data processing patterns, Alink can be a valuable option for developers looking for robust big data and advanced machine learning tools,” said Jia Yangqing, president and senior fellow of Data Platform at Alibaba Cloud Intelligence.

The core codes of Alink have been made available on GitHub, the world’s largest developer community. For Alibaba, opening-source its algorithm platform will potentially bring new and faster updates from global developers.

Jia, who served Facebook before joining Alibaba, said that Alibaba Cloud is committed to connecting with the open-source community “as early as possible” for software development, and added that his company is already one of the top ten contributors to the GitHub community.

Pinduoduo, one of Alibaba’s rising rivals, has already developed a machine learning system to display and recommend products to shoppers before they even know they need them.

Alibaba’s recommendation system is now search-based, while Pinduoduo’s recommendation system has counted more factors including social elements. For example, most of Pinduoduo’s users like to share product links on WeChat to their friends to form a team and to get a discount. Their interactions could help Pinduoduo’s recommendation system to predict what other goods might be needed and what might not be demanded, based on the recorded data.