Baidu follows US tech giants and open sources its deep learning tools

Baidu follows US tech giants and open sources its deep learning tools ,

Sean Gallup/Getty Images

Microsoft, Google, Facebook, and Amazon have all done it — and now Baidu's doing it, too. The Chinese tech giant has open sourced one of its key machine learning tools, PaddlePaddle, offering the software up to the global community of AI researchers.

This move has become common among tech firms as they pour more and more resources into their AI work. Open sourcing your tools is a good way to attract talent, but it also allows companies to shape the development of a field that's becoming increasingly central to consumer tech, underpinning everything from voice interfaces to auto-sorting photo galleries.

Baidu's big claim for PaddlePaddle is that it's easier to use than rival programs. Like Amazon's DSSTNE and Microsoft's CNTK, PaddlePaddle offers a toolkit for deep learning, but Baidu says comparable software is designed to work in too many different situations, making it less approachable to newcomers. Xu Wei, the leader of Baidu's PaddlePaddle development, tells The Verge that a machine translation program written with Baidu's software needs only a quarter of the amount of code demanded by other deep learning tools.

"PaddlePaddle provides a simple on-ramp to deep learning."

"In order to achieve wider adoption of this technology, engineers, developers and even coding enthusiasts need to have an easy to use framework," Xu told The Verge over email. "PaddlePaddle provides a simple on-ramp to deep learning, while also providing [...] scalability." A pre-alpha version is available on GitHub now, with a full release scheduled for September 30th.

Baidu is hoping this ease of use will make PaddlePaddle more attractive to computer scientists, and draw attention away from machine learning tools released by Google and Facebook. Google, especially is the leader in open-source AI tools, at least by one measure of success — the number of interactions with the program at code repository GitHub. As noted by Quartz recently, Google's TensorFlow is significantly more popular Facebook's Torch toolkit.

Xu told The Verge that if any one company's machine learning tools became too dominant it would be like "trying to use one programming language to code all applications." Xu's belief is that no one company will dominate this area. "Different tools have different strengths," he says. "The deep learning ecosystem will end up having different tools optimized for different uses. Just like no programing language truly dominates software development."

Baidu says PaddlePaddle is already being used by more than 30 of its offline and online products and services, covering sectors from search to finance to health. It even helps work out how long food will take to reach customers in Baidu's takeout business. Baidu says its deep learning tools can guarantee customers a 50 percent refund if deliveries are not made at the correct time, and seems just as confident about the utility of its AI toolkit.