On Wednesday at the Baidu World stage in Beijing, Chinese internet giant Baidu announced its plan to offer an open source AI software platform for enterprises.
Baidu, the "Google of China," has been working to establish itself as a leader in creating deep learning software. The platform, called "PaddlePaddle," was originally developed for internal use at Baidu as an integral piece in search ranking, advertising, image classification, translation and driverless cars.
Andrew Ng, Baidu's chief scientist and a leading researcher in machine learning, believes that making Baidu's platform open-source will help companies who want to get started in machine learning.
SEE: Machine learning: The smart person's guide (TechRepublic)
"We're giving enterprises the easy on-ramp they need to make AI a reality," Ng said in a statement. "Baidu is going to release the AI technology that powers Baidu's offerings as open source offerings that they feel will be the easiest to use and most accessible for all developers, not just researchers and data scientists."
Getting started in AI, Ng said, can often be expensive or burdensome, and Baidu's "intuitive" platform is intended to offer "a natural starting point for programmers and enthusiasts who want to apply deep learning to their projects and products."
The platform will support neural network architectures like convolutional neural networks and recurrent neural networks. It is also efficient, optimizing "mathematical operations via state-of- the-art BLAS libraries, e.g. Intel MKL, ATLAS, OpenBLAS and cuBLAS, etc." And PaddlePaddle is scalable, as well, available to be used on "large numbers of GPUs/CPUs on multiple machines," according to Baidu's release.
In order to get started with PaddlePaddle, you can check out its user guide, available in English and Chinese, which has examples and demos in image classification, natural language processing and recommendations.
SEE: Machine automation policy guidelines template (Tech Pro Research)
The move seems to be part of Baidu's broader AI strategy. Back in January, it began testing autonomous driving technology in China. And it also announced this week that it would be conducting autonomous research in California, alongside Tesla, Ford, and Google, to develop a self-driving car, which it plans to unveil in the next five years.
It's worth noting that Baidu is not the first to make its AI platform available. In January, Microsoft offered open source machine learning on GitHub. Google has its own deep learning platform, TensorFlow. In May, Amazon announced its own Deep Scalable Sparse Tensor Network Engine (DSSTNE) library. And there are a handful of others, as well, like Caffe, Theano and Torch.
SEE: How Google is getting smarter with artificial intelligence (CBS News)
Still, Baidu believes it is poised to offer something the other platforms don't have—an easy tool for those new to AI.
"Baidu is not the first to go open source with their software for deep learning," said Roman Yampolskiy, director of the Cybersecurity lab at the University of Louisville. "The goal, of course, is to become a default platform for research and development, in particular for less technically savvy developers."
- Amazon open sources its deep learning software (ZDNet)
- How Amazon wants to bridge the data science gap by bringing machine learning to the cloud (TechRepublic)
- Machine learning face-off: Microsoft uses Band to show what its Watson rival is capable of (ZDNet)
- Google DeepMind: The smart person's guide (TechRepublic)
- How developers can take advantage of machine learning on Google Cloud Platform (TechRepublic)
Hope Reese has nothing to disclose. She doesn't hold investments in the technology companies she covers.
Hope Reese is a journalist in Louisville, KY. Her writing has been featured in The Atlantic, The Boston Globe, The Chicago Tribune, Playboy, Undark Magazine, VICE, Vox, and other publications.