Chinese search giant Baidu has decided to open source its deep learning software to the greater community of AI software developers.
Baidu claims the software, dubbed “PaddlePaddle”, is superior to its competitors as it can be easily used by software developers without much experience.
“Other deep learning platforms have been a great boon to researchers wanting to invent new deep learning algorithms, but their high degree of flexibility limits their ease of use,” said Andrew Ng, chief scientist at Baidu Research and founder of deep learning research lab Google Brain.
“In contrast, Paddle focuses on making it easy for enthusiasts and programmers – not just machine learning researchers – to learn and use powerful deep learning tools.”
According to Xu, PaddlePaddle exposes only high-level functions to developers, removing the need for developers to have complex mathematical knowledge in order to use the platform. It also uses significantly less code than others of its kind.
“Take machine translation model as an example,” said Xu in a comment to InfoWorld.
“You only need to write 25 per cent of code [compared to] what you would on some other popular platforms. PaddlePaddle allows you to apply existing models to new problems without worrying [about] the math equations used to implement the model.”
While Baidu has released a number of other open source projects, this is the first machine learning project offered by the company, and the only one with documentation and examples in both English and Chinese.
The company will release the software with documentation and specifications to GitHub on September 30 under an Apache open source license. However, a pre-release Alpha version is already available for developers.